DESC:RESERVATION OF RIGHTS
[001] A portion of the disclosure of this patent document contains material, which is subject to intellectual property rights such as, but are not limited to, copyright, design, trademark, Integrated Circuit (IC) layout design, and/or trade dress protection, belonging to Jio Platforms Limited (JPL) or its affiliates (hereinafter referred as owner). The owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent files or records, but otherwise reserves all rights whatsoever. All rights to such intellectual property are fully reserved by the owner.

FIELD OF DISCLOSURE
[002] The embodiments of the present disclosure generally relate to story-level recommendations of magazine articles on a magazine aggregator website. In particular, the present disclosure relates to a system and a method for retrieving stories from a digital magazine and providing personalized recommendations and the stories to a user.

BACKGROUND OF DISCLOSURE
[003] The following description of related art is intended to provide background information pertaining to the field of the disclosure. This section may include certain aspects of the art that may be related to various features of the present disclosure. However, it should be appreciated that this section be used only to enhance the understanding of the reader with respect to the present disclosure, and not as admissions of prior art.
[004] Generally, magazine aggregators receive magazines from multiple publishers as a single document containing multiple stories. However, users may prefer only a few stories out of all the stories in magazines. So, it is important for providing recommendations and retrieving individual stories from digital magazines. The retrieval of stories from magazines is challenging because the layout of magazines is different for different publishers. There is no straightforward way to define article boundaries across the magazines with different layouts. Further, each magazine article may consist of several different types of pages. There may be index pages, advertisement pages, pages with photographs, and the like. The appearance of these pages is quite inter-mixed across magazines.
[005] Typically, as index pages of several different magazines vary and as the index pages cannot be relied on for determining stories boundaries, accurate detection of stories boundaries through the index pages and indexes is difficult. Further, advertisements that are embedded with news articles may create a false sense of completion of the story but may or may not be correlated to the story itself and hence may make detection of stories boundaries cumbersome and inaccurate. Furthermore, in cases when there are multiple stories on a single magazine page, detection of boundaries of stories becomes a tedious task.
[006] There is, therefore, a need in the art to improve detection of individual stories from the digital magazines for retrieval and recommendation to users to overcome the deficiencies of the prior arts.

OBJECTS OF THE PRESENT DISCLOSURE
[007] Some of the objects of the present disclosure, which at least one embodiment herein satisfies are as listed herein below.
[008] It is an object of the present disclosure to provide a system and a method for retrieving stories from a digital magazine.
[009] It is an object of the present disclosure to extract and ensemble page layout features, text-based features, and image-based features from each page of a plurality of pages of the digital magazine.
[0010] It is an object of the present disclosure to classify the plurality of pages of the digital magazine into a plurality of page types and identify page type for each page of the plurality of pages of the digital magazine.
[0011] It is an object of the present disclosure to segment and retrieve stories from the plurality of pages based on the ensembled features and the page types.
[0012] It is an object of the present disclosure to provide personalised magazine story recommendations to a user for better user engagement and satisfaction.

SUMMARY
[0013] This section is provided to introduce certain objects and aspects of the present disclosure in a simplified form that are further described below in the detailed description. This summary is not intended to identify the key features or the scope of the claimed subject matter.
[0014] In an aspect, the present disclosure relates to a system for providing personalized recommendations to a user. The system includes one or more processors, and a memory operatively coupled to the one or more processors. The memory includes processor-executable instructions, which on execution, cause the one or more processors to extract page layout features, text-based features, and image-based features from each page of a plurality of pages of the digital magazine, ensemble the page layout features, the text-based features, and the image-based features to classify the plurality of pages of the digital magazine into a plurality of page types, identify at least one page type for each page of the plurality of pages to segment stories from the plurality of pages based on the ensembled features, and retrieve stories from the plurality of pages of the digital magazine and provide personalized recommendations and stories to the user based on the identified at least one page type.
[0015] In an embodiment, the page layout features may include at least one of: a font of words, a colour of the words, a number of lines in the at least one page, a number of camel case words in at least one page, and a ratio of bold words in to normal words in the at least one page.
[0016] In an embodiment, the text-based features may include at least one of: one or more texts, a number of words in at least one page, a number of capital words in the at least one page, a number of question words in at least one page, a number of dates mentioned in at least one page, a number of phone numbers mentioned in at least one page, a number of emails mentioned in the at least one page, a count of numbers in at least one page, a count of punctuations in at least one page, and a ratio of capital to total words in at least one page.
[0017] In an embodiment, the one or more processors may extract the text-based features by cleaning, normalising, and tokenizing the one or more texts.
[0018] In an embodiment, the memory includes processor-executable instructions, which on execution, may cause the one or more processors to identify, via a Named Entity Recognition (NER) model, at least one entity from the tokenized one or more texts.
[0019] In an embodiment, the memory includes processor-executable instructions, which on execution, may cause the one or more processors to obtain feature vectors for the one or more texts to determine a similarity score between the one or more texts.
[0020] In an embodiment, the image-based features may include at least one of: a number of images in the at least one page, a ratio of an area occupied by the images in the at least one page, a number of objects available in the images, a dimension of the images in the at least one page, and a position of the images in the at least one page.
[0021] In an embodiment, the one or more processors may classify the plurality of pages of the digital magazine into the plurality of page types by being configured to compare the page layout features, the text-based features, and the image-based features extracted from one page of the plurality of pages with the page layout features, the text-based features, and the image-based features extracted from at least one another page of the plurality of pages, identify similarity and linkage between at least two pages of the plurality of pages, and classify the plurality of pages into the plurality of page types based on the identification of the similarity and the linkage.
[0022] In an embodiment, the at least one page type may include at least one of a cover page, an index page, a story beginning page, a story end page, multi-story pages, and advertisement pages.
[0023] In an embodiment, the memory includes processor-executable instructions, which on execution, may cause the one or more processors to delete at least one duplicate page from the plurality of pages based on the comparison between the page layout features, the text-based features, and the image-based features extracted from one page of the plurality of pages and the page layout features, the text-based features, and the image-based features extracted from at least one another page of the plurality of pages.
[0024] In an embodiment, the one or more processors may segment the stories from the plurality of pages by being configured to: detect story boundary pages from the plurality of pages based on the identified at least one page type, determine a coherence score for the story boundary pages by extracting the at least one entity identified from the tokenized one or more texts for all pages based on the ensembled features, determine a probability score for the story boundary pages based on the identified at least one page type, and segment the stories from the plurality of pages based on the coherence score and the probability score of the story boundary pages.
[0025] In an embodiment, the one or more processors may retrieve the stories from the plurality of pages of the digital magazine by being configured to obtain a story boundary from the plurality of pages by aggregating the coherence score and the probability score of the story boundary pages, and retrieve the stories from the plurality of pages of the digital magazine based on the story boundary.
[0026] In another aspect, the present disclosure relates to a method for providing personalized recommendations to a user. The method includes extracting, by a processor associated with a system, page layout features, text-based features, and image-based features from each page of a plurality of pages of the digital magazine, ensembling, by the processor, the page layout features, the text-based features, and the image-based features to classify the plurality of pages of the digital magazine into a plurality of page types, identifying, by the processor, at least one page type for each page of the plurality of pages to segment stories from the plurality of pages based on the ensembled features, and retrieving, by the processor, the stories from the plurality of pages of the digital magazine and providing personalized recommendations and the stories to the user based on the identified at least one page type.
[0027] In an embodiment, extracting, by the processor, the text-based features may include cleaning, normalising, and tokenizing one or more texts.
[0028] In an embodiment, the method may include identifying, via a Named Entity Recognition (NER) model, by the processor, at least one entity from the tokenized one or more texts.
[0029] In an embodiment, the method may include obtaining, by the processor, feature vectors for the one or more texts to determine a similarity score between the one or more texts.
[0030] In an embodiment, classifying, by the processor, the plurality of pages of the digital magazine into the plurality of page types may include comparing, by the processor, the page layout features, the text-based features, and the image-based features extracted from one page of the plurality of pages with the page layout features, the text-based features, and the image-based features extracted from at least one another page of the plurality of pages, identifying, by the processor, similarity and linkage between at least two pages of the plurality of pages, and classifying, by the processor, the plurality of pages into the plurality of page types based on an identification of the similarity and the linkage.
[0031] In an embodiment, the method may include deleting, by the processor, at least one duplicate page from the plurality of pages based on a comparison between the page layout features, the text-based features, and the image-based features extracted from one page of the plurality of pages and the page layout features, the text-based features, and the image-based features extracted from at least one another page of the plurality of pages.
[0032] In an embodiment, segmenting, by the processor, the stories from the plurality of pages may include detecting, by the processor, story boundary pages from the plurality of pages based on the identified at least one page type, determining, by the processor, a coherence score for the story boundary pages by extracting the at least one entity identified from tokenized one or more texts for all pages based on the ensembled features, determining, by the processor, a probability score for the story boundary pages based on the identified at least one page type, and segmenting, by the processor, the stories from the plurality of pages based on the coherence score and the probability score of the story boundary pages.
[0033] In an embodiment, retrieving, by the processor, the stories from the plurality of pages of the digital magazine may include obtaining, by the processor, a story boundary from the plurality of pages by aggregating the coherence score and the probability score of the story boundary pages, and retrieving, by the processor, the stories from the plurality of pages of the digital magazine based on the story boundary.
[0034] In another aspect, the present disclosure relates to a user equipment. The user equipment includes one or more processors, and a memory operatively coupled to the one or more processors, wherein the memory includes processor-executable instructions, which on execution, cause the one or more processors to receive one or more personalized recommendations and stories from a system. The one or more processors are communicatively coupled with the system. The system is configured to extract page layout features, text-based features, and image-based features from each page of a plurality of pages of a digital magazine, ensemble the page layout features, the text-based features, and the image-based features to classify the plurality of pages of the digital magazine into a plurality of page types, identify at least one page type for each page of the plurality of pages to segment stories from the plurality of pages based on the ensembled features, and retrieve the stories from the plurality of pages of the digital magazine and provide the one or more personalized recommendations and the stories to the user equipment based on the identified at least one page type.

BRIEF DESCRIPTION OF THE DRAWINGS
[0035] The accompanying drawings, which are incorporated herein, and constitute a part of this invention, illustrate exemplary embodiments of the disclosed methods and systems in which like reference numerals refer to the same parts throughout the different drawings. Components in the drawings are not necessarily to scale, emphasis instead being placed upon clearly illustrating the principles of the present invention. Some drawings may indicate the components using block diagrams and may not represent the internal circuitry of each component. It will be appreciated by those skilled in the art that invention of such drawings includes the invention of electrical components, electronic components or circuitry commonly used to implement such components.
[0036] The diagrams are for illustration only, which thus is not a limitation of the present disclosure, and wherein:
[0037] FIG. 1A illustrates an exemplary block diagram (100A) of a stories retrieval and recommendation system (110), in accordance with an embodiment of the present disclosure.
[0038] FIG. 1B illustrates an exemplary network architecture (100B) in which or with which embodiments of the present disclosure may be implemented.
[0039] FIG. 2 illustrates an exemplary flow diagram (200) for retrieval of stories from a digital magazine, in accordance with an embodiment of the present disclosure.
[0040] FIG. 3 illustrates an exemplary flow diagram (300) for training a random forest model for page type classification, in accordance with an embodiment of the present disclosure.
[0041] FIG. 4 illustrates an exemplary flow diagram (400) of page type classification inference, in accordance with an embodiment of the present disclosure.
[0042] FIG. 5 illustrates an exemplary flow diagram (500) for creation of a Latent Semantic Indexing (LSI) model, in accordance with an embodiment of the present disclosure.
[0043] FIG. 6 illustrates an exemplary flow diagram (600) for layout feature extraction mechanism, in accordance with an embodiment of the present disclosure.
[0044] FIG. 7 illustrates an exemplary flow diagram (700) for text-based feature extraction mechanism, in accordance with an embodiment of the present disclosure.
[0045] FIG. 8 illustrates an exemplary flow diagram (800) for image-based feature extraction, in accordance with an embodiment of the present disclosure.
[0046] FIG. 9 illustrates an exemplary flow diagram (900) for story boundary detection using a text tilling mechanism, in accordance with an embodiment of the present disclosure.
[0047] FIG. 10 illustrates an exemplary computer system (1000) in which or with which embodiments of the present invention can be utilized in accordance with embodiments of the present disclosure.
[0048] The foregoing shall be more apparent from the following more detailed description of the disclosure.

DETAILED DESCRIPTION
[0049] In the following description, for the purposes of explanation, various specific details are set forth in order to provide a thorough understanding of embodiments of the present disclosure. It will be apparent, however, that embodiments of the present disclosure may be practiced without these specific details. Several features described hereafter can each be used independently of one another or with any combination of other features. An individual feature may not address all of the problems discussed above or might address only some of the problems discussed above. Some of the problems discussed above might not be fully addressed by any of the features described herein.
[0050] The ensuing description provides exemplary embodiments only, and is not intended to limit the scope, applicability, or configuration of the disclosure. Rather, the ensuing description of the exemplary embodiments will provide those skilled in the art with an enabling description for implementing an exemplary embodiment. It should be understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope of the disclosure as set forth.
[0051] Specific details are given in the following description to provide a thorough understanding of the embodiments. However, it will be understood by one of ordinary skill in the art that the embodiments may be practiced without these specific details. For example, circuits, systems, networks, processes, and other components may be shown as components in block diagram form in order not to obscure the embodiments in unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail in order to avoid obscuring the embodiments.
[0052] Also, it is noted that individual embodiments may be described as a process which is depicted as a flowchart, a flow diagram, a data flow diagram, a structure diagram, or a block diagram. Although a flowchart may describe the operations as a sequential process, many of the operations can be performed in parallel or concurrently. In addition, the order of the operations may be re-arranged. A process is terminated when its operations are completed but could have additional steps not included in a figure. A process may correspond to a method, a function, a procedure, a subroutine, a subprogram, etc. When a process corresponds to a function, its termination can correspond to a return of the function to the calling function or the main function.
[0053] The word “exemplary” and/or “demonstrative” is used herein to mean serving as an example, instance, or illustration. For the avoidance of doubt, the subject matter disclosed herein is not limited by such examples. In addition, any aspect or design described herein as “exemplary” and/or “demonstrative” is not necessarily to be construed as preferred or advantageous over other aspects or designs, nor is it meant to preclude equivalent exemplary structures and techniques known to those of ordinary skill in the art. Furthermore, to the extent that the terms “includes,” “has,” “contains,” and other similar words are used in either the detailed description or the claims, such terms are intended to be inclusive in a manner similar to the term “comprising” as an open transition word without precluding any additional or other elements.
[0054] Reference throughout this specification to “one embodiment” or “an embodiment” or “an instance” or “one instance” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Thus, the appearances of the phrases “in one embodiment” or “in an embodiment” in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
[0055] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises” and/or “comprising,” when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. As used herein, the term “and/or” includes any and all combinations of one or more of the associated listed items.
[0056] The present disclosure describes a lightweight and highly accurate technique of segmenting magazines and retrieving story boundaries from magazines. The present disclosure performs layout features extraction, text-based features extraction, images and image properties extraction from multiple pages, inter-page linking and comparison, magazine pages classification, and story indexes retrieval from index pages. Further, the proposed disclosure performs duplicate pages detection across the magazines and segregates the magazine into the stories efficiently.
[0057] Therefore, embodiments of the present disclosure relate to effectively providing product named personalized magazine story level recommendations in a magazine aggregator website and providing personalized stories from various magazines which are relevant to users.
[0058] Various embodiments of the present disclosure will be explained in detail with reference to FIGs. 1A-10.
[0059] FIG. 1A illustrates an exemplary block diagram (100A) of a stories retrieval and recommendation system (110), in accordance with an embodiment of the present disclosure.
[0060] In an embodiment, and as shown in FIG. 1A, the system (110) may include one or more processors (102). The one or more processors (102) may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, logic circuitries, and/or any devices that manipulate data based on operational instructions. Among other capabilities, the one or more processor(s) (102) may be configured to fetch and execute computer-readable instructions stored in a memory (104) of the system (110). The memory (104) may store one or more computer-readable instructions or routines, which may be fetched and executed to create or share the data units over a network service. The memory (104) may comprise any non-transitory storage device including, for example, volatile memory such as Random Access Memory (RAM), or non-volatile memory such as Erasable Programmable Read-Only Memory (EPROM), flash memory, and the like.
[0061] The system (110) may also comprise an interface(s) (106). The interface(s) (106) may comprise a variety of interfaces, for example, interfaces for data input and output devices, referred to as I/O devices, storage devices, and the like. The interface(s) (106) may facilitate communication of the system (110) with various devices coupled to it. The interface(s) (106) may also provide a communication pathway for one or more components of a processing engine (108). Examples of such components include, but are not limited to, processing engine(s) (108) and a database (130).
[0062] The processing engine(s) (108) may be implemented as a combination of hardware and programming (for example, programmable instructions) to implement one or more functionalities of the processing engine(s) (108). In examples described herein, such combinations of hardware and programming may be implemented in several different ways. For example, the programming for the processing engine(s) (108) may be processor executable instructions stored on a non-transitory machine-readable storage medium and the hardware for the one or more processors (102) may comprise a processing resource (for example, one or more processors), to execute such instructions. In the present examples, the machine-readable storage medium may store instructions that, when executed by the processing resource, implement the processing engine(s) (108). In such examples, the system (110) may comprise the machine-readable storage medium storing the instructions and the processing resource to execute the instructions, or the machine-readable storage medium may be separate but accessible to the system (110) and the processing resource. In other examples, the processing engine(s) (108) may be implemented by electronic circuitry.
[0063] In an aspect, the database (130) may comprise data that may be either stored or generated as a result of functionalities implemented by any of the components of the processor(s) (102) or the processing engine(s) (108) or the system (110). In an exemplary embodiment, the processing engine(s) (108) may include a layout features extraction engine (112), a text-based features extraction engine (114), an image-based features extraction engine (116), an inter-page linking and comparison engine (118), a magazine pages classification engine (120), a story indexes retrieval engine (122), a duplicate pages detection engine (124), a segregation engine (126), and other engine(s) (128), wherein the other engines (128) may further include, without limitation, data receiving engine, storage engine, computing engine, or signal generation engine. Other engine(s) (128) may supplement the functionalities of the processing engine(s) (108) or the system (110). The system (110) may be implemented using any or a combination of hardware components and software components.
[0064] In an embodiment, the layout features extraction engine (112) may extract page layout features, for example, but not limited to, text, font, and colour from each page of the plurality of pages of the magazines using a standard Optical Character Recognition (OCR) technique. Fonts of different types and colour may be considered during extraction of the text.
[0065] In an embodiment, the text-based features extraction engine (114) may extract text-based features from each page of the plurality of pages of the magazines and create advanced features using the extracted text. The text-based features extraction engine (114) may use a topic modelling method at various granularities (e.g., global, magazine, or category level) to create topics in an unsupervised manner. Further, a text tiling method may be used to calculate the features on basis of similarity between neighbourhood pages. In addition, an entities extraction may be performed using the extracted text.
[0066] In an embodiment, the image-based features extraction engine (116) may extract image-based features, for example, but not limited to, images and image properties from each page of the plurality of pages of the magazines, using, for example, person detection and recognition in the images, object detection and recognition in the images, dimensions of the images and position of the images on the page.
[0067] In an embodiment, the page layout features, the text-based features, and the image-based features may be ensembled to classify the plurality of pages of the magazine into a plurality of page types.
[0068] In an embodiment, the inter-page linking and comparison engine (118) may determine similarity of entities and linkage between the pages of the magazines using a knowledge graph. A comparison mechanism may be used to compare page level entities extracted from multiple pages where the entities may be determined based on a similarity on context of predicate.
[0069] In an embodiment, the magazine pages classification engine (120) may classify magazine pages into one of types of pages, for example, a cover page, an index page, story beginning page, story end page, multi-story pages, and advertisement pages, based on the ensembled features. The magazine pages classification engine (120) may identify a page type for each page of the plurality of pages, to segment stories from the plurality of pages based on the ensembled features. The magazine pages classification may be performed by using text and image features retrieved using multiple processes.
[0070] In an embodiment, the story indexes retrieval engine (122) may retrieve story indexes from the index page after classification of the magazine pages. The story indexes retrieval engine (122) may retrieve the stories from the plurality of pages of the magazine and provide personalized recommendations and the stories to the user.
[0071] In an embodiment, the duplicate pages detection engine (124) may remove the duplicate pages from the determined set of pages.
[0072] In an embodiment, the segregation engine (126) may provide a mechanism that segregates the magazine into one or more stories.
[0073] In an embodiment, the present disclosure may provide Portable Document Format (PDF) files of magazine as an input to the system (110). The system (110) may extract a variety of features viz. page layout features, text-based features, and image-based features. By way of an example, different types of features may be:
(a) number of images in a page [Image-based],
(b) a ratio of area occupied by images in a page [Image-based],
(c) a number of lines in a page [Layout Features],
(d) a number of words in a page [Text-based],
(e) a number of camel case words in a page [Layout Features],
(f) a number of capital words in a page [Text-based],
(g) a number of question words (what, how, when etc.) in a page [Text-based],
(h) a number of dates mentioned in a page [Text-based],
(i) a number of phone numbers mentioned in a page [Text-based],
(j) a number of emails mentioned in a page [Text-based],
(k) count of numbers in a page [Text-based],
(l) count of punctuations in a page [Text-based],
(m) a ratio of bold words in a page to normal words in a page [Layout features], and
(n) ratio capital to total words in a page [Text-based].
[0074] Each of the above-mentioned exemplary features may be used for magazine page classification. After page type classification, the above-mentioned exemplary features along with other features like topic models, text tiling algorithms, and knowledge graph may be used to perform stories segmentation.
[0075] Although FIG. 1 shows an exemplary block diagram (100) of a stories retrieval and recommendation system (110), in other embodiments, the stories retrieval and recommendation system (110) may include fewer components, different components, differently arranged components, or additional functional components than depicted in FIG. 1. Additionally, or alternatively, one or more components of the stories retrieval and recommendation system (110) may perform functions described as being performed by one or more other components of the stories retrieval and recommendation system (110).
[0076] FIG. 1B illustrates an exemplary network architecture (100B) in which or with which embodiments of the present disclosure may be implemented.
[0077] Referring to FIG. 1B, the network architecture (100B) may include one or more user equipments (134-1, 134-2…134-N) associated with one or more users (132-1, 132-2…132-N) in an environment. A person of ordinary skill in the art will understand that one or more users (132-1, 132-2…132-N) may be individually referred to as the user (132) and collectively referred to as the users (132). Similarly, a person of ordinary skill in the art will understand that one or more user equipments (134-1, 134-2…134-N) may be individually referred to as the user equipment (134) and collectively referred to as the user equipment (134). A person of ordinary skill in the art will appreciate that the terms “computing device(s)” and “user equipment” may be used interchangeably throughout the disclosure. Although three user equipments (134) are depicted in FIG. 11, however any number of the user equipments (134) may be included without departing from the scope of the ongoing description.
[0078] In an embodiment, the user equipment (134) may include smart devices operating in a smart environment, for example, an Internet of Things (IoT) system. In such an embodiment, the user equipment (134) may include, but is not limited to, smart phones, smart watches, smart sensors (e.g., mechanical, thermal, electrical, magnetic, etc.), networked appliances, networked peripheral devices, networked lighting system, communication devices, networked vehicle accessories, networked vehicular devices, smart accessories, tablets, smart television (TV), computers, smart security system, smart home system, other devices for monitoring or interacting with or for the users (132) and/or entities, or any combination thereof.
[0079] A person of ordinary skill in the art will appreciate that the user equipment (134) may include, but is not limited to, intelligent, multi-sensing, network-connected devices, that can integrate seamlessly with each other and/or with a central server or a cloud-computing system or any other device that is network-connected.
[0080] In an embodiment, the user equipment (134) may include, but is not limited to, a handheld wireless communication device (e.g., a mobile phone, a smart phone, a phablet device, and so on), a wearable computer device (e.g., a head-mounted display computer device, a head-mounted camera device, a wristwatch computer device, and so on), a Global Positioning System (GPS) device, a laptop computer, a tablet computer, or another type of portable computer, a media playing device, a portable gaming system, and/or any other type of computer device with wireless communication capabilities, and the like. In an embodiment, the user equipment (134) may include, but is not limited to, any electrical, electronic, electro-mechanical, or an equipment, or a combination of one or more of the above devices such as virtual reality (VR) devices, augmented reality (AR) devices, laptop, a general-purpose computer, desktop, personal digital assistant, tablet computer, mainframe computer, or any other computing device, wherein the user equipment (134) may include one or more in-built or externally coupled accessories including, but not limited to, a visual aid device such as a camera, an audio aid, a microphone, a keyboard, and input devices for receiving input from the user or the entity such as touch pad, touch enabled screen, electronic pen, and the like.
[0081] A person of ordinary skill in the art will appreciate that the user equipment (134) may not be restricted to the mentioned devices and various other devices may be used.
[0082] Referring to FIG. 1B, the user equipment (134) may communicate with a system (110), for example, the stories retrieval and recommendation system, through a network (136). In an embodiment, the network (136) may include at least one of a Fifth Generation (5G) network, or the like. The network (136) may enable the user equipment (134) to communicate with other devices in the network architecture (100B) and/or with the system (110). The network (136) may include a wireless card or some other transceiver connection to facilitate this communication. In another embodiment, the network (136) may be implemented as, or include any of a variety of different communication technologies such as a wide area network (WAN), a local area network (LAN), a wireless network, a mobile network, a Virtual Private Network (VPN), the Internet, the Public Switched Telephone Network (PSTN), or the like.
[0083] In accordance with embodiments of the present disclosure, the system (110) may be designed and configured for extracting page layout features, text-based features, and image-based features from each page of the plurality of pages of the digital magazine. As such, the system (110) may ensemble the page layout features, the text-based features, and the image-based features to classify the plurality of pages of the digital magazine into a plurality of page types. The system (110) may identify at least one page type for each page of the plurality of pages to segment stories from the plurality of pages based on the ensembled features. Alternatively, or additionally, the system (110) may retrieve the stories from the plurality of pages of the digital magazine and provide personalized recommendations and the stories to the user based on the ensembled features and the identified page type.
[0084] Although FIG. 1B shows exemplary components of the network architecture (100B), in other embodiments, the network architecture (100B) may include fewer components, different components, differently arranged components, or additional functional components than depicted in FIG. 1B. Additionally, or alternatively, one or more components of the network architecture (100B) may perform functions described as being performed by one or more other components of the network architecture (100B).
[0085] FIG. 2 illustrates an exemplary flow diagram (200) for retrieval of stories from a digital magazine (202), in accordance with an embodiment of the present disclosure.
[0086] With respect to FIG. 2, the digital magazine (202) may be segmented into multiple pages at (204). The segmented magazine pages may be used for page level features extraction (206). The page level feature extraction (206) may involve layout features extraction, text-based features extraction, and image-based features extraction. The extracted page level features (208) may be used to perform page-type classification (210). The page-type classification (210) may be performed to determine page type(s) probability (212). Further, an index page may be identified at (214) and probable index pages with probabilities may be determined at (216) using the page type(s) probability (212). In response to an identification of the index page at (214) and a determination of the probable index pages with probabilities at (216), a story start page number may be predicted at (218) and subsequently story start pages basis index page(s) and probabilities may be determined at (220).
[0087] In an embodiment, the segmented magazines pages may be used to extract entities from pages at (222). At 224, the pages with extracted entities may be used as input for determining knowledge graph-based page coherence. Further, the knowledge graph-based page coherence may be determined at (226).
[0088] In an embodiment, a text tiling method may be used for text-based feature extraction (228) based on the extracted page level features (208). The text-based features obtained by the text tiling method (228) and the output obtained from the knowledge graph-based page coherence at (226) may be used to determine consecutive pages and coherence score at (230).
[0089] In an embodiment, the consecutive pages and coherence score determined at (230) may be used to detect story boundary at (232) based on an input received from the determined page type(s) probability (212) and the determined story start pages basis index page(s) and probabilities (220). Upon detection of the story boundary detection at (232), an output with respect to multiple stories, for example, story 1, story 2,…., story N may be retrieved at (234).
[0090] In an embodiment, for recommendations of stories from the digital magazines at personalized level, a user profile may be created in terms of pages and stories read by the user and a matching method may be used to retrieve relevant story for the user.
[0091] FIG. 3 illustrates an exemplary flow diagram (300) for training a random forest model for page type classification, in accordance with an embodiment of the present disclosure.
[0092] With reference to FIG. 3, magazines corpus (302) may be used and segmented into multiple pages at (304). The magazine pages (306) may be used for page level features extraction i.e., layout features extraction, text-based features extraction, and image-based feature extraction at (308). The extracted page level features may be pre-processed at (310) and a feature selection may be performed at (312). The selected features may be sent as an input to train a random forest model (314). Further, the random forest model (314) may be trained and saved at (316) for future use and reference.
[0093] FIG. 4 illustrates an exemplary flow diagram (400) of page type classification inference, in accordance with an embodiment of the present disclosure.
[0094] With reference to FIG. 4, the magazine may be segmented into multiple pages at (402). The segmented magazine pages (404) may be used for page level feature extraction at (406). At (408), the page level extracted features may be pre-processed to perform page type classification. Further, at (410), the page type classification may be performed using a random forest model to determine page type(s) probability. The determined page type(s) probability may be saved at (412).
[0095] FIG. 5 illustrates an exemplary flow diagram (500) for creation of a Latent Semantic Indexing (LSI) model, in accordance with an embodiment of the present disclosure.
[0096] With reference to FIG. 5, the magazine corpus (302) may be used to extract text from magazines at (502). Cleaning and normalisation of the extracted text may be performed at (504). At (506), the cleansed and normalised text may be used to train the LSI model. Further, the trained LSI model may be saved at (508).
[0097] FIG. 6 illustrates an exemplary flow diagram (600) for layout feature extraction mechanism, in accordance with an embodiment of the present disclosure.
[0098] With respect to FIG. 6, the magazine may be read using a PDF reader or an OCR model at (602). At (604), the layout features may be extracted from the magazine. The layout features may include, but not limited to, a number of lines in a page of the magazine, a number of camel case words in the page of the magazine, a determination of whether the word is bold or normal, a ratio of bold word in the page to normal words in the page of the magazine, a font of the words, a colour of the words, and the like.
[0099] FIG. 7 illustrates an exemplary flow diagram (700) for text-based feature extraction mechanism, in accordance with an embodiment of the present disclosure.
[00100] With respect to FIG. 7, a PDF reader/OCR technique may be used to read the magazine at (702). At (704), relevant texts may be extracted from the magazine. Further, text-based features extraction may be performed at (706) using three different approaches. In a first approach, features such as a number of words in the page, a number of capital words in the page, a number of question words (what, how, when etc.) in the page, a number of dates mentioned in the page, a number of phone numbers mentioned in the page, and the like may be extracted by processing the text extracted using the PDF reader or the OCR model.
[00101] In a second approach, entities level features may be extracted by following steps: (a) cleaning and normalization of the extracted text may be performed at (708), and (b) a tokenisation of the extracted text may be performed at (710). A Named Entity Recognition (NER) model may be trained at (712) and a LSI model may be obtained at (714) using the cleaned, normalized and tokenised text. By using the NER model, the entities may be extracted from the tokenized text at (716). In a third approach, the LSI model may be used to obtain LSI feature vector at (718) for the tokenized text, which may be further utilized to compute the similarity score between the texts.
[00102] FIG. 8 illustrates an exemplary flow diagram (800) for image-based feature extraction, in accordance with an embodiment of the present disclosure.
[00103] With respect to FIG. 8, at (802), PDF reader/OCR technique may be used to read the magazine. The image-based features may be extracted, at (804), from the magazines read using the PDF reader/OCR techniques. The image-based features such as a number of images in the page, a ratio of area occupied by the images in the page, a position of image on the page and a dimension of image may be extracted. The extracted image-based features may be identified as images at (806).
[00104] In an embodiment, deep learning-based approaches such as an object detection model (808) and a face detection and recognition model (810) may be used for extraction of the image-based features.
[00105] In an embodiment, the object detection model (808) may be implemented to extract information about an object present in the image, for example, how many and what are the objects present in the image.
[00106] In an embodiment, the face detection and recognition model (810) may be used to identify the number of persons present in the image that can provide more information about the image as well as the magazine page to identify each person.
[00107] FIG. 9 illustrates an exemplary flow diagram (900) for story boundary detection using a text tilling mechanism, in accordance with an embodiment of the present disclosure.
[00108] With respect to FIG. 9, the page level features extraction may be performed at (902) to train a random forest model. The trained random forest model may be loaded at (904) to perform page type classification. The page type classification may be performed, at (906), using the trained random forest model. In response to performing the page type classification, story pages with probability may be extracted at (908). The story pages with probability may be used to train an LSI model and the trained LSI model may be loaded at (910). The trained LSI model may be loaded to obtain LSI vector for each page at (914). The LSI vector may be used to determine cosine similarity between the LSI vectors at (918). The cosine similarity between the LSI vectors may be used for story boundary detection using the text tilling method at (922). Further, the text tilling method may be used to obtain story boundary pages with probable score at (924).
[00109] In an embodiment, the story pages at (908) may be used to extract entities for all pages at (912). The extracted entities may be used to train knowledge graph model and the trained knowledge graph model may be loaded at (916). The trained knowledge graph model may be used to determine knowledge graph-based coherence store at (920). The determined knowledge graph-based coherence store may be used to obtain the story boundary pages with coherence score (926). Further, the probable score obtained at (924) and the coherence score obtained at (928) may be aggregated to obtain a story boundary with aggregated score. The story boundary with aggregated score may be saved at (930) for future use and reference.
[00110] The present disclosure may improve business significantly as users/consumers appreciate to receive the stories of their choice rather than the entire magazine in recommendations.
[00111] FIG. 10 illustrates an exemplary computer system (1000) in which or with which embodiments of the present disclosure may be implemented.
[00112] As shown in FIG. 10, the computer system (1000) may include an external storage device (1010), a bus (1020), a main memory (1030), a read only memory (1040), a mass storage device (1050), a communication port (1060), and a processor (1070).
[00113] A person skilled in the art will appreciate that the computer system (1000) may include more than one processor and communication ports. The processor (1070) may include various modules associated with embodiments of the present disclosure.
[00114] In an embodiment, the communication port (1060) may be any of an RS-232 port for use with a modem-based dialup connection, a 10/100 Ethernet port, a Gigabit or 10 Gigabit port using copper or fiber, a serial port, a parallel port, or other existing or future ports. The communication port (1060) may be chosen depending on a network, such a Local Area Network (LAN), Wide Area Network (WAN), or any network to which the computer system (1600) connects.
[00115] In an embodiment, the memory (1030) may be a Random Access Memory (RAM), or any other dynamic storage device commonly known in the art. The read-only memory (1040) may be any static storage device(s) e.g., but not limited to, a Programmable Read Only Memory (PROM) chips for storing static information e.g., start-up or Basic Input/Output system (BIOS) instructions for the processor (1070).
[00116] In an embodiment, the mass storage (1050) may be any current or future mass storage solution, which can be used to store information and/or instructions. Exemplary mass storage solutions include, but are not limited to, Parallel Advanced Technology Attachment (PATA) or Serial Advanced Technology Attachment (SATA) hard disk drives or solid-state drives (internal or external, e.g., having Universal Serial Bus (USB) and/or Firewire interfaces), one or more optical discs, Redundant Array of Independent Disks (RAID) storage, e.g., an array of disks (e.g., SATA arrays).
[00117] In an embodiment, the bus (1020) may communicatively couple the processor(s) (1070) with the other memory, storage, and communication blocks. The bus (1020) may be, e.g., a Peripheral Component Interconnect (PCI)/PCI Extended (PCI-X) bus, Small Computer System Interface (SCSI), Universal Serial Bus (USB) or the like, for connecting expansion cards, drives, and other subsystems as well as other buses, such a front side bus (FSB), which connects the processor (1070) to the computer system (1000).
[00118] Optionally, operator and administrative interfaces, e.g., a display, keyboard, joystick, and a cursor control device, may also be coupled to the bus (1020) to support direct operator interaction with the computer system (1000). Other operator and administrative interfaces may be provided through network connections connected through the communication port (1060). Components described above may be meant only to exemplify various possibilities. In no way should the aforementioned exemplary computer system (1000) limit the scope of the present disclosure.
[00119] While the foregoing describes various embodiments of the invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof. The scope of the present disclosure is determined by the claims that follow. The present disclosure may not be limited to the described embodiments, versions or examples, which are included to enable a person having ordinary skill in the art to make and use the present disclosure when combined with information and knowledge available to the person having ordinary skill in the art.

ADVANTAGES OF THE PRESENT DISCLOSURE
[00120] The present disclosure efficiently retrieves stories from a digital magazine.
[00121] The present disclosure extracts and ensembles page layout features, text-based features, and image-based features from each page of the digital magazine.
[00122] The present disclosure classifies the plurality of pages of the digital magazine into a plurality of page types.
[00123] The present disclosure identifies page type for each page of the plurality of pages of the digital magazine.
[00124] The present disclosure segments and retrieves stories from the plurality of pages of the digital magazine based on the ensembled features and the page types.
[00125] The present disclosure provides personalised magazine story recommendations to a user for better user engagement and satisfaction.
,CLAIMS:1. A system (110) for providing personalized recommendations to a user, the system (110) comprising:
one or more processors (102); and
a memory (104) operatively coupled to the one or more processors (102), wherein the memory (104) comprises processor-executable instructions, which on execution, cause the one or more processors (102) to:
extract page layout features, text-based features, and image-based features from each page of a plurality of pages of the digital magazine;
ensemble the page layout features, the text-based features, and the image-based features to classify the plurality of pages of the digital magazine into a plurality of page types;
identify at least one page type for each page of the plurality of pages to segment stories from the plurality of pages based on the ensembled features; and
retrieve the stories from the plurality of pages of the digital magazine and provide personalized recommendations and the stories to the user based on the identified at least one page type.

2. The system (110) as claimed in claim 1, wherein the page layout features comprise at least one of: a font of words, a colour of the words, a number of lines in the at least one page, a number of camel case words in the at least one page, and a ratio of bold words to normal words in the at least one page.

3. The system (110) as claimed in claim 1, wherein the text-based features comprise at least one of: one or more texts, a number of words in the at least one page, a number of capital words in the at least one page, a number of question words in the at least one page, a number of dates mentioned in the at least one page, a number of phone numbers mentioned in the at least one page, a number of emails mentioned in the at least one page, a count of numbers in the at least one page, a count of punctuations in the at least one page, and a ratio of capital to total words in the at least one page.

4. The system (110) as claimed in claim 3, wherein the one or more processors (102) are to extract the text-based features by cleaning, normalising, and tokenizing the one or more texts.

5. The system (110) as claimed in claim 4, wherein the memory (104) comprises processor-executable instructions, which on execution, cause the one or more processors (102) to identify, via a Named Entity Recognition (NER) model, at least one entity from the tokenized one or more texts.

6. The system (110) as claimed in claim 3, wherein the memory (104) comprises processor-executable instructions, which on execution, cause the one or more processors (102) to obtain feature vectors for the one or more texts to determine a similarity score between the one or more texts.

7. The system (110) as claimed in claim 1, wherein the image-based features comprise at least one of: a number of images in the at least one page, a ratio of an area occupied by the images in the at least one page, a number of objects available in the images, a dimension of the images in the at least one page, and a position of the images in the at least one page.

8. The system (110) as claimed in claim 1, wherein the one or more processors (102) are to classify the plurality of pages of the digital magazine into the plurality of page types by being configured to:
compare the page layout features, the text-based features, and the image-based features extracted from one page of the plurality of pages with the page layout features, the text-based features, and the image-based features extracted from at least one another page of the plurality of pages;
identify similarity and linkage between at least two pages of the plurality of pages based on the comparison; and
classify the plurality of pages into the plurality of page types based on the identification of the similarity and the linkage.

9. The system (110) as claimed in claim 1, wherein the at least one page type comprises at least one of: a cover page, an index page, a story beginning page, a story end page, multi-story pages, and advertisement pages.

10. The system (110) as claimed in claim 8, wherein the memory (104) comprises processor-executable instructions, which on execution, cause the one or more processors (102) to delete at least one duplicate page from the plurality of pages based on the comparison between the page layout features, the text-based features, and the image-based features extracted from one page of the plurality of pages and the page layout features, the text-based features, and the image-based features extracted from at least one another page of the plurality of pages.

11. The system (110) as claimed in claim 5, wherein the one or more processors (102) are to segment the stories from the plurality of pages by being configured to:
detect story boundary pages from the plurality of pages based on the identified at least one page type;
determine a coherence score for the story boundary pages by extracting the at least one entity identified from the tokenized one or more texts for all pages based on the ensembled features;
determine a probability score for the story boundary pages based on the identified at least one page type; and
segment the stories from the plurality of pages based on the coherence score and the probability score of the story boundary pages.

12. The system (110) as claimed in claim 11, wherein the one or more processors (102) are to retrieve the stories from the plurality of pages of the digital magazine by being configured to:
obtain a story boundary from the plurality of pages by aggregating the coherence score and the probability score of the story boundary pages; and
retrieve the stories from the plurality of pages of the digital magazine based on the story boundary.

13. A method for providing personalized recommendations to a user, the method comprising:
extracting, by a processor (102) associated with a system (110), page layout features, text-based features and image-based features from each page of a plurality of pages of the digital magazine;
ensembling, by the processor (102), the page layout features, the text-based features and the image-based features to classify the plurality of pages of the digital magazine into a plurality of page types;
identifying, by the processor (102), at least one page type for each page of the plurality of pages to segment stories from the plurality of pages based on the ensembled features; and
retrieving, by the processor (102), the stories from the plurality of pages of the digital magazine, and providing personalized recommendations and the stories to the user based on the identified at least one page type.

14. The method as claimed in claim 13, wherein extracting, by the processor (102), the text-based features comprise cleaning, normalising and tokenizing one or more texts.

15. The method as claimed in claim 14, comprising identifying, via a Named Entity Recognition (NER) model, by the processor (102), at least one entity from the tokenized one or more texts.

16. The method as claimed in claim 13, comprising obtaining, by the processor (102), feature vectors for one or more texts to determine a similarity score between the one or more texts.

17. The method as claimed in claim 13, wherein classifying, by the processor (102), the plurality of pages of the digital magazine into the plurality of page types comprises:
comparing, by the processor (102), the page layout features, the text-based features, and the image-based features extracted from one page of the plurality of pages with the page layout features, the text-based features, and the image-based features extracted from at least one another page of the plurality of pages;
identifying, by the processor (102), similarity and linkage between at least two pages of the plurality of pages based on the comparison; and
classifying, by the processor (102), the plurality of pages into the plurality of page types based on the identification of the similarity and the linkage.

18. The method as claimed in claim 17, comprising deleting, by the processor (102), at least one duplicate page from the plurality of pages based on the comparison between the page layout features, the text-based features, and the image-based features extracted from one page of the plurality of pages and the page layout features, the text-based features, and the image-based features extracted from at least one another page of the plurality of pages.

19. The method as claimed in claim 15, wherein segmenting, by the processor (102), the stories from the plurality of pages comprises:
detecting, by the processor (102), story boundary pages from the plurality of pages based on the identified at least one page type;
determining, by the processor (102), a coherence score for the story boundary pages by extracting the at least one entity identified from the tokenized one or more texts for all pages based on the ensembled features;
determining, by the processor (102), a probability score for the story boundary pages based on the identified at least one page type; and
segmenting, by the processor (102), the stories from the plurality of pages based on the coherence score and the probability score of the story boundary pages.

20. The method as claimed in claim 19, wherein retrieving, by the processor (102), the stories from the plurality of pages of the digital magazine comprises:
obtaining, by the processor (102), a story boundary from the plurality of pages by aggregating the coherence score and the probability score of the story boundary pages; and
retrieving, by the processor (102), the stories from the plurality of pages of the digital magazine based on the story boundary.

21. A user equipment, comprising:
one or more processors; and
a memory operatively coupled to the one or more processors, wherein the memory comprises processor-executable instructions, which on execution, cause the one or more processors to:
receive one or more personalized recommendations and stories from a system (110),
wherein the one or more processors are communicatively coupled with the system (110), and wherein the system (110) is configured to:
extract page layout features, text-based features, and image-based features from each page of a plurality of pages of a digital magazine;
ensemble the page layout features, the text-based features, and the image-based features to classify the plurality of pages of the digital magazine into a plurality of page types;
identify at least one page type for each page of the plurality of pages to segment stories from the plurality of pages based on the ensembled features; and
retrieve the stories from the plurality of pages of the digital magazine and provide the one or more personalized recommendations and the stories to the user equipment based on the identified at least one page type.