FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION (See section 10, rule 13)
1. Title of the invention: ESTIMATION OF APPEARANCE INDEX OF TYRES
2. Applicant(s)
NAME NATIONALITY ADDRESS
CEAT LIMITED Indian CEAT Ltd At: Get Muwala Po: Chandrapura Ta: Halol - 389 350 Dist: Panchmahal, Gujarat, India
3. Preamble to the description
COMPLETE SPECIFICATION
The following specification particularly describes the invention and the manner in which it
is to be performed.

TECHNICAL FIELD
[0001] The present subject matter relates, in general, to estimation of
appearance index of tyres. In particular, the present subject matter relates to methods and devices for appearance index estimation of passenger car, truck, and bus tyres.
BACKGROUND
[0002] Passenger Car Radial tyres also known as PCR tyres, and Truck Bus
Radial tires also known as TBR tyres, may be subject to wear and tear which may be caused due to various reasons such as misalignment of tires, pressure levels, external influences, incorrect mounting, braking impact, and the like. An analysis of the various defects that may be caused is performed in order to determine an appearance index. Determination of the appearance index helps in development of new products or releasing original equipment approvals and the like. Generally, an analysis of the defects on the tire is performed by a subjective approach which requires high expertise.
BRIEF DESCRIPTION OF DRAWINGS
[0003] The features, aspects, and advantages of the present subject matter will
be better understood with the help of the following description and accompanying
figures. The use of the same reference number in different figures indicates similar
or identical features and components.
[0004] Fig. 1 illustrates a block diagram of an appearance index estimation
system, in accordance with an example of the present subject matter.
[0005] Fig. 2(a) illustrates a chipping defect in a passenger car tyre, in
accordance with an example of the present subject matter.
[0006] Fig. 2(b) illustrates a chunking defect in a passenger car tyre, in
accordance with an example of the present subject matter.
[0007] Fig. 2(c) illustrates an abrasion defect in the passenger car tyre, in
accordance with an example of the present subject matter.

[0008] Fig. 3 illustrates an example output of the appearance index estimation
system, in accordance with an example of the present subject matter.
[0009] Figs. 4 illustrates a method for appearance index estimation, in
accordance with an example of the present subject matter.
DETAILED DESCRIPTION
[0010] The present subject matter relates to appearance index estimation and,
specifically to methods and devices for appearance index estimation of passenger
car and truck, bus tyres which provide an accurate and relevant appearance index.
[0011] In known techniques, estimation of appearance index is highly
dependent on personnel with high expertise to manually study the tyre and then determine the appearance index. In order to overcome the challenges with respect to conventional techniques, the present subject matter provides techniques for accurate estimation of appearance index of a tire. In operation, an image of a tire is received, where the image includes at least one defect. The image of the tire is analyzed to identify the at least one defect. Each of the identified at least one defect is categorized into any one of a plurality of pre-determined categories. On categorizing each of the at least one defect, an extent of degradation of the tire is computed. The extent of degradation is computed based on a number of defects identified in the image. Based on the extent of degradation of the tire and the pre-determined category each of the at least one defect belongs to an appearance index is estimated.
[0012] Therefore, techniques of the present subject matter provide an
improved accuracy in appearance index estimation with an increased speed of estimation without professional intervention. Further, techniques of the present subject matter utilize machine learning based objective rating in order to provide real time responses, with minimal dependency on human expertise.
[0013] The above and other features, aspects, and advantages of the subject
matter will be better explained with regard to the following description and accompanying figures. It should be noted that the description and figures merely illustrate the principles of the present subject matter along with examples described

herein and should not be construed as a limitation to the present subject matter. It is thus understood that various arrangements may be devised that, although not explicitly described or shown herein, embody the principles of the present disclosure. Moreover, all statements herein reciting principles, aspects, and examples thereof, are intended to encompass equivalents thereof. Further, for the sake of simplicity, and without limitation, the same numbers are used throughout the drawings to reference like features and components.
[0014] Fig. 1 illustrates a block diagram of an appearance index estimation
system 100, in accordance with an example implementation of the present subject
matter. In one example, the appearance index estimation system 100 may be part of
a source device (not shown in the figure). The source device may be an Internet of
things (IOT) device, a computing device, a personal computer, a laptop, a tablet, a
mobile phone, and the like. In another example, the appearance index estimation
system 100, alternatively referred to as system 100, may be hosted on a server (not
shown in the figure) that may communicate with the source device. In one example,
the system 100 may communicatively be coupled to a plurality of user devices 102-
1, 102-2, …, 102-n, alternatively and collectively referred to as user device 102.
[0015] For example, the user device 102 may be a mobile phone, or any
device with which the user may be able to capture an image to be shared with the system 100. In one example, the system 100 and the plurality of user devices 102 may communicate over a network 104. The network 104 may be a wireless network or a combination of a wired and wireless network. The network 104 can also include a collection of individual networks, interconnected with each other and functioning as a single large network, such as the Internet. Examples of such individual networks include, but are not limited to, Global System for Mobile Communication (GSM) network, Universal Mobile Telecommunications System (UMTS) network, Personal Communications Service (PCS) network, Time Division Multiple Access (TDMA) network, Code Division Multiple Access (CDMA) network, Next Generation Network (NGN), Public Switched Telephone Network (PSTN), Long Term Evolution (LTE), and Integrated Services Digital Network (ISDN).

[0016] The appearance index estimation system 100 may include a processor
106 and a memory 108 coupled to the processor 106. The functions of functional block labelled as “processor(s)”, may be provided through the use of dedicated hardware as well as hardware capable of executing instructions. When provided by a processor, the functions may be provided by a single dedicated processor, by a single shared processor, or by a plurality of individual processors, some of which may be shared. Moreover, explicit use of the term “processor” would not be construed to refer exclusively to hardware capable of executing instructions, and may implicitly include, without limitation, digital signal processor (DSP) hardware, network processor, application specific integrated circuit (ASIC), field programmable gate array (FPGA), read only memory (ROM) for storing instructions, random access memory (RAM), non-volatile storage. Other hardware, standard and/or custom, may also be included.
[0017] The memory 108 may include any computer-readable medium
including, for example, volatile memory (e.g., RAM), and/or non-volatile memory (e.g., EPROM, flash memory, etc.).
[0018] The appearance index estimation system 100 may further include
engines 110, such as an analyzing engine 112 and a deep learning engine 114. In one example, the engines 110 may be implemented as a combination of hardware and firmware. In examples described herein, such combinations of hardware and firmware may be implemented in several different ways. For example, the firmware for the module may be processor executable instructions stored on a non-transitory machine-readable storage medium and the hardware for the module may include a processing resource (for example, implemented as either a single processor or a combination of multiple processors), to execute such instructions.
[0019] In the present examples, the machine-readable storage medium may
store instructions that, when executed by the processing resource, implement the functionalities of the engines. In such examples, the appearance index estimation system 100 may include the machine-readable storage medium storing the instructions and the processing resource to execute the instructions. In other examples of the present subject matter, the machine-readable storage medium may

be located at a different location but accessible to the appearance index estimation system 100 and the processor 106.
[0020] In one example, the appearance index estimation system 100 may also
include interfaces 116 which may include a variety of computer-readable instructions-based interfaces and hardware interfaces that allow interaction with other communication, storage, and computing devices, such as network entities, web servers, databases, and external repositories, and peripheral devices. In one example, interfaces 116 may be used to view the results obtained from the deep learning engines and inputs received from the plurality of user devices 102. For example, the inputs obtained from the user devices 102 and the results obtained from the deep learning engine 114 may be viewed on a graphical user interface like a display screen.
[0021] The appearance index estimation system 100 may further include data
118, that serves, amongst other things, as a repository for storing data that may be fetched, processed, received, or generated by the analyzing engine 112 and the deep learning engine 114. The data 118 may include communication data, user inputs, calibration parameters, and the like. In an example, the data 118 may be stored in the memory 108. Although the following description has been predominantly explained with respect to the appearance index estimation system 100 communicating with a single user device 102, similar principles may be applicable to the plurality of user devices 102 communicating with the appearance index estimation system 100.
[0022] In one example, the appearance index estimation system 100 may be
trained to estimate an appearance index of a tire. In one example, the tire may be a passenger car tyre. In another example, the tire may be a truck tire or a bus tire. In one example, in order to train the system 100, a plurality of images of a tire with various defects may be provided. The plurality images may be obtained from different tires exposed to different scenarios, where the images may include diverse forms of defects on the tires to train the system 100 for better accuracy of identification and detection. In one example, pre-processing techniques, such as

blurring the image, rotating images, edge detections, and the like may be performed
in order to train the model for the diverse input that may be received in real time.
[0023] On receiving the plurality of images, the images may be analyzed for
different types of defects. In one example, each defect may be categorized into a plurality of pre-determined categories. For example, a chipping defect, a chunking defect, and an abrasion defect may be considered. The chipping of the tyre may be a resultant of a cut caused by the effect of tractive braking and causing tearing of the rubber compound, usually at an angle of 90° to the direction of the cut. Chunking of the tyre is the tearing away of large chips and abrasion is a resultant of fatigue cut growth and the peeling off of the rubber debris. Although the following description has been explained with chipping, chunking, and abrasion as the three defects identified, similar principles are applicable to any other type of defects on the tire.
[0024] On identification of the defect and a number of each of the defects, a
bounding box may be drawn to indicate a type of the defect identified and location of the defect with different annotations. For example, an image of the tyre received for a tire from a user may include multiple chips, chunks, and/or abrasions. A bounding box corresponding to each defect may be drawn to annotate each such defect. For instance, a first bounding box with a first annotation may be drawn to represent a chip, a second bounding box with a second annotation may be drawn to represent a chunk, and a third bounding box with a third annotation may be drawn to represent an abrasion. In one example, annotating the defects may be done by personnel with high expertise. In an example where the defects may be overlapping one another, the boundary boxes drawn to annotate the defects may also overlap one another.
[0025] In this manner, each defect identified may be categorized into any one
of the predefined categories. In one example, a pre-defined category may be indicative of the type of defect identified. For example, a first pre-determined category may be a chipping category, a second pre-determined category may be a chunking category, and a third pre-determined category may be an abrasion category. Although the following description describes chipping, chunking, and

abrasion as the three types of pre-determined categories based on the type of defects identified with the help of the following example, other methods of categorization and categories may also be applicable for the system 100 in accordance with the principles of the present subject matter.
[0026] On categorizing the defects, an extent of degradation of the tire may
be computed. In one example, the extent of degradation may be computed based on
a number of defects detected and identified. The extent of degradation of the tire
may be compared with a benchmark tire in order to be categorized into any one of
the following classes as acceptable, borderline acceptable, or unacceptable. Based
on the extent of degradation of the tire and a number of defects categorized into the
pre-determined categories indicated by bounding boxes, a rating score may be
assigned against each of the image analysed to train the deep learning model 114.
[0027] The images along with the extent of the degradation of the tire, the
number of defects identified, a size of the bounding box, and the number of bounding boxes may be input to the deep learning engine 114 of the system 100. In one example, the deep learning engine may be a Convoluted Neural Network (CNN) backbone architecture, understood to a person skilled in the art. In one example, the input images may be passed through a plurality of channels of the CNN backbone. Techniques such as feature extraction, regression, and the like may be performed to identify the defects and create bounding boxes for each of the defects identified.
[0028] During the training phase, the deep learning engine 114 may be
configured to identify the defects in the image and apply methods of defect detection and recognition on the input data in order to create boundary boxes on the defects identified. The boundary boxes that are created by the engine 114 may be an output of the training along with the appearance index estimation. In one example, the number of boundary boxes output by the system 100 may be based on a number of defects in the image of the tire. Based on the number of bounding boxes, and the extent of degradation of the tire, the appearance index may be estimated. In one example, the predicted output may be compared to the input provided to the deep learning engine in order to determine an accuracy of the

training. In a scenario where the predicted output is distant to the input provided, a feedback mechanism may be provided to the engine in order to increase the accuracy of prediction. In a scenario where the prediction is inaccurate, the model may be re-trained.
[0029] On deployment of the system 100, in one example, the system 100
may receive an input image of a tyre with at least one defect. In one example, the system 100 may receive the input image from a user. In another example, the system 100 may receive the input image from another system or device. In one example, the input image may be a photograph taken by the user. The photograph may be taken based on a predefined criteria.
[0030] On receiving the input image, the analyzing engine 112 of the system
100 may be configured to analyse the input image to identify the at least one defect
and categorize each of the at least one defect into any one of a pre-determined
category. For example, if the image from the user has more than one defect, such
as a chip and a chunk, the system 100 may be configured to analyse the image to
identify each of the defect and categorize the defect into a pre-determined category.
[0031] On identifying each of the at least one defect in the image provided by
the user, the deep learning engine 114 of the system 100 may be configured to create a bounding box for each defect identified. In one example, the deep learning engine 114 may be a Convoluted Neural Network (CNN) backbone architecture, understood to a person skilled in the art. In one example, the input images may be passed through a plurality of channels of the CNN backbone, to extract the features of the image.
[0032] Once the features are extracted, the features may be classified based
on methods known in the art. On categorizing the features, bounding boxes may be created based on the types of defects, with methods of detection and recognition. Taking the same example as explained above, if the image has a chip and a chunk, a first bounding box with a first annotation may be created to represent the chip and a location of the chip. Similarly, a second bounding box with a second annotation may be created to represent the chunk. In a scenario where the image includes more than one chip, or more than one chunk, at different locations, a bounding box with

the first annotation may be created for each chip at each location and a bounding box with the second annotation may be created for each chunk at each location. In one example, principles of non-maximum suppression may be incorporated in order to eliminate unnecessary bounding boxes and to identify bounding boxes that are relevant to the input.
[0033] On creating the bounding boxes, an extent of degradation of the tire
may be computed. In one example, the extent of the degradation may be calculated based on a number of defects identified The extent of the degradation may be categorized into any one of the following classes of acceptable, borderline acceptable, and unacceptable. Based on the total number of bounding boxes and the extent of the degradation of the tire the appearance index of the tire may be estimated. Similar to the extent of the defect, the appearance index may also be classified as acceptable, borderline acceptable, and unacceptable. Interfaces of the system 100 may display the appearance index, the number of bounding boxes, a location of the bounding boxes, and the like to the user.
[0034] Therefore, techniques of the present subject matter provide improved
speed and precision in estimation of the appearance index of a tyre without human intervention.
[0035] Fig. 2(a) illustrates a chipping defect in a passenger car tyre, in
accordance with an example of the present subject matter. In one example, a
passenger car tyre 202 is depicted in the figure, with a chipping defect 204. In one
example the chipping defect 204 may be categorized as the first pre-determined
category as discussed above. Similarly, Fig. 2(b) illustrates a chunking defect 206
in a passenger car tyre 202, in accordance with an example of the present subject
matter, where the chunking defect may be categorized as the second pre-determined
category and Fig. 2(c) illustrates an abrasion defect 208 in the passenger car tyre
202, in accordance with an example of the present subject matter, where the
abrasion defect may be categorized as the third pre-determined category.
[0036] Fig. 3 illustrates an example output of the appearance index estimation
system, in accordance with an example of the present subject matter. An example output image 302 of a passenger car tyre 304 is depicted where, a first bounding

box 306 with a first annotation may be created to represent chipping defect. Similarly, a second bounding box 308 with a second annotation may be created to represent chunking defect and a third bounding box 310 may be created with a third annotation to represent the abrasion defect. As can be observed from the figure, the example output image includes two bounding boxes with the first annotation to represent chipping, two bounding boxes with the second annotation to represent chunking, and one bounding box with the third bounding box to represent an abrasion. Based on these number of bounding boxes as represented and the extent of degradation of the tire, the appearance index may be estimated and categorized into any one of acceptable, borderline acceptable, or unacceptable.
[0037] Fig. 4 illustrates a method for appearance index estimation 400, in
accordance with an example of the present subject matter. The order in which the method 400 is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement method 400 or an alternative method. Additionally, individual blocks may be deleted from the method 400 without departing from the spirit and scope of the subject matter described herein. Furthermore, the method 400 may be implemented in any suitable hardware, computer readable instructions, firmware, or combination thereof. For discussion, the method 400 is described with reference to the implementations illustrated in Fig(s). 1-Fig. 3.
[0038] At block 402 of the method 400, an image of a tire is received, wherein
the image comprises at least one defect. In one example, a user may provide the image of the tire.
[0039] At block 404 of the method 400, the image of the tire is analyzed to
identify the at least one defect. In one example, but not limited to the at least one defect may be any one of a chipping defect, a chunking defect, an abrasion defect, and the like.
[0040] At block 406 of the method 400, each of the identified at least one
defect is categorized into any one of a plurality of pre-determined categories. In one example, the predefined categories may be any one of a cut, a chunk, and an abrasion.

[0041] At block 408 of the method 400, an extent of degradation the tire is
computed based on a number of defects identified in the image.
[0042] At block 410 of the method 400, an appearance index is estimated
based on the extent of degradation of the tire and the pre-determined category each of the at least one defect belongs to. In one example, the appearance index is classified into any one of an acceptable class, a borderline acceptable class, or an unacceptable class.
[0043] Although the present subject matter has been described with reference
to specific embodiments, this description is not meant to be construed in a limiting sense. Various modifications of the disclosed embodiments, as well as alternate embodiments of the subject matter, will become apparent to persons skilled in the art upon reference to the description of the subject matter.

I/We Claim:
1. A method for appearance index estimation, the method comprising:
receiving an image of a tire, wherein the image comprises at least one defect;
analyzing the image of the tire to identify the at least one defect;
categorizing each of the identified at least one defect into any one of a plurality of pre-determined categories;
computing an extent of degradation of the tire based on a number of defects identified in the image; and
estimating an appearance index based on the extent of degradation of the tire and the pre-determined category each of the at least one defect belongs to.
2. The method as claimed in claim 1 further comprising creating a boundary box for each of the at least one defect based on a location of the at least one defect.
3. The method as claimed in claim 2 further comprising:
calculating a number of boundary boxes created for each category of the plurality of pre-defined categories; and
displaying the number of boundary boxes created.
4. The method as claimed in claim 1, wherein the predefined categories are any one of a cut, a chunk, and an abrasion.
5. The method as claimed in claim 1, wherein the extent of degradation of the tire is calculated based a comparison between the number of defects on the tire and a defect of a benchmark tire.

6. The method as claimed in claim 1 further comprises training a deep learning engine by providing an input image to the deep learning engine, wherein the input image comprises a bounding box with a corresponding annotation created around each defect of the at least one defect.
7. The method as claimed in claim 1, wherein the appearance index is classified into any one of an acceptable class, a borderline acceptable class, or an unacceptable class.
8. An appearance index estimation system comprising:
a deep learning engine configured to:
obtain an image of a tire from a user, wherein the image of the tire comprises at least one defect;
analyze the image of the tire to identify the at least one defect;
categorize the at least one defect identified into any one of a plurality of pre-determined categories; and
compute an extent of degradation of the tire based on a number of defects identified in the image; and
determine an appearance index based on the extent of degradation of the tire and the pre-determined category the at least one defect belongs to.
9. The appearance index estimation system as claimed in claim 8, wherein the deep learning engine is to further create a boundary box for each category of the plurality of pre-defined categories based on a location of the at least one defect.
10. The appearance index estimation system as claimed in claim 8, wherein the deep learning engine is further to:

calculate a number of boundary boxes created for each category of the plurality of pre-defined categories; and
display the number of boundary boxes created.