DESC:A REAL-TIME SAMPLE CLASSIFIER
FIELD OF INVENTION
The invention generally relates to the field of image processing
and particularly to a method and a system for real-time classification of a sample.
BACKGROUND
Digital pathology involves generating digital slide scans from
glass slides using a scanning device. Digital pathology allows for
high resolution viewing, managing, analysing and interpreting
digital information. The technique used in digital pathology includes but is not limited to whole slide scanning, virtual microscopy and image processing. There are various digital pathology systems available in the art which allows for scanning of slides, uploading of scanned images, automated analysis of images, classification of images based on classification algorithms and exchange of images over a computer network. One such system discloses a database comprising images and quantitative information on diagnostically relevant features, an inspection procedure for the examination of a suspected object and analysis tools for comparison of the locally obtained data to those stored in the database.
Another system allows capturing of images, performing image
analysis for identifying abnormal and normal segments, calculating deviation between normal and abnormal segments and ranking objects based on calculated deviations. The detection and identification of pathologies is also done remotely via internet.
Yet another such system provides a shared memory configured to store scanned images, an execution controller configured to determine potential regions of interest within the scanned image and to divide the sample image into sub-sections, one or more processors configured to analyze the potential regions of interest
to produce intermediate results and to further analyze the sub-sections responsive to the user’s request from a client device to produce a final analysis based on the intermediate results and a network transceiver configured to transmit the sub-sections to the client device and to transmit the final analysis to the client device.
One significant disadvantage of the above mentioned systems is that the classification algorithms cannot be iterated and upgraded on real-time basis; as a result, the chances of errors in classifying cells in an image are more.

BRIEF DESCRIPTION OF DRAWINGS
So that the manner in which the recited features of the invention can be understood in detail, some of the embodiments are illustrated in the appended drawings. It is to be noted, however, that the appended drawings illustrate only typical embodiments of this invention and are therefore not to be considered limiting of its scope, for the invention may admit to other equally effective embodiments.
FIG.1 shows a flow chart depicting a method for real-time classification of a sample, according to an embodiment of the invention.
FIG. 2 shows a schematic representation of the system for real-time classification of a sample, according to an embodiment of the invention.
FIGS 3a – 3n generally show representative screen shots of the user interface, according to an embodiment of the invention.

SUMMARY OF THE INVENTION
One aspect of the invention provides a method for real-time classification of a biological sample. The method includes selecting an image of the biological sample. A preliminary classification of the image is obtained. The preliminary classification is then analysed to determine the accuracy of the classification. The analysed classification is then subjected to a plurality of iterations to obtain a final classification.
Another aspect of the invention provides a system for real-time classification of the sample. The system includes a plurality of source devices for generating a scanned image of the sample. A storage unit for retrievably storing the images is connected to each of the source devices. An analyser for classification of the stored images is connected to the storage means. A display unit is connected to the analyser. An interactive user interface is enabled on the display unit for analysing and viewing images.

DETAILED DESCRIPTION OF THE INVENTION
Various embodiments of the invention provide a method and a system for real-time classification of a biological sample. The invention provides a method for real-time classification of a biological sample. The method includes selecting an image of the biological sample. A preliminary classification of the image is obtained. The preliminary classification is then analysed to determine the accuracy of the classification. The analysed classification is then subjected to iterations to obtain a final classification. The method described briefly herein above shall be described in detail.
The method includes selecting a scanned image of a biological sample by a first user. The first user can be an unskilled person or a skilled person. Examples of the biological sample include but are not limited to a cell, a tissue, a smear of thin blood film, a thick blood film, a Pap smear, sputum, a biological fluid, a cerebrospinal fluid, a fluid extracted from inflammatory site, a fluid extracted from tumor site or a fluid containing pathogens.
The scanning of the sample is done with the help of a scanning device. Examples of scanning devices used for the purpose of the present invention include but are not limited to a whole slide scanner, a scanning microscope, an imaging flow cytometer, an optofluidic microscope, a slide profiler and other such devices as known to a person skilled in the art. Alternatively, the image to be classified can be a pre-scanned image already obtained from a set of stored images.
Once the image is selected, the preliminary classification of the images is achieved through a manual process or an automated process. The classification through automated process is achieved using a classification algorithm. The cells in the scanned image are classified by the classification algorithm as normal, abnormal or suspected to be abnormal to generate a preliminary classification. Examples of classification algorithm include but are not limited to a genetic algorithm, a machine learning algorithm, an NLP algorithm, a neural network based processing approach, and all such algorithms capable of handling complex data and analysis. Further, the cells in the scanned images can also be classified as belonging to a particular class or a cell type. The example of class includes but is not limited to a healthy Red Blood Cells (RBC), a malaria infected RBCs (iRBC), platelets, and White Blood Cells (WBC), in the case of blood smear profiling for quantitative diagnosis of malaria ; a monocyte, a lymphocyte, a neutrophil, a basophil and an eosinophil in case of classifying WBC for Complete Blood Count. The images of the cells classified into respective identified classes are then displayed as preliminary classification. In one example of the invention, the time taken to obtain an automated preliminary classification is in the range of about 10seconds to about 50seconds.
In one embodiment of the invention, the first user performs the preliminary classification manually. Subsequent to obtaining the preliminary classification, the first user submits the preliminary classification to a classification engine to obtain a final classification. The final classification obtained is then cross validated by the first user.
In another example of the invention, the first user submits the preliminary classification for review to a second user. Subsequent to review of the preliminary classification by the second user, a final classification report is generated. The final classification report is then sent to the first user.
In yet another example of the invention, the first user submits the preliminary classification obtained for review to a plurality of second users. The preliminary classification obtained can be generated by the first user. Alternatively, the preliminary classification obtained can be generated by the classification algorithm. The second users selected can be an expert in the subject matter related to the sample being classified. Each of the second users generates a second classification report. Each of the second classification reports generated by each of the second users is assigned a score. The scores are then collated to obtain a list of weighted average scores generated in respect of each of the classified image. A ranking of the scores is then obtained to select the highest ranking image. The classification report with respect to the highest ranking image is then sent to the first user.
The invention provides a system for the real-time classification of the sample. The system includes a plurality of source devices for generating a scanned image of the sample. A storage unit for retrievably storing the images is connected to each of the source devices. An analyser for classification of the stored images is connected to the storage unit. A display unit is connected to the analyser. An interactive user interface is enabled on a display system for analysing and viewing images. The system described herein above briefly shall be described in detail.
FIG. 2 shows a schematic representation of the system for real-time classification of a sample, according to an embodiment of the invention. The system includes a plurality of source devices 201 having means of generating a plurality of scanned images of the sample corresponding to the source device 201. Scanning of the sample is achieved by a scanning device. The scanning devices include but are not limited to a whole slide scanner, a scanning microscope, an imaging flow cytometer, an opto-fluidic microscope, a slide profiler and other such devices as known to a person skilled in the art. The source devices 201 are remotely connected to a storage unit 203 for storing the scanned images. The remote connection is achieved through a network. The network described herein includes but is not limited to a local area network (LAN), a wide area network (WAN), a wireless network, a satellite network or a combination thereof. The scanned images obtained from the scanning device are uploaded and retrievably stored. The examples of storage unit includes but are not limited to a resident server, a remote server, a web server, a virtual server, example a cloud or a combination thereof. The analyzer 205 classifies the stored scanned images using classification algorithms. In one embodiment of the system, the analyzer can be a remote unit or a centrally located. Alternatively, the analyzer 205 can be a unit locally connected to the source device 201.
The invention further provides an interactive user interface for real time classification of the biological sample. The interface includes a user authentication. A menu for selecting an image is also provided. The user is provided with an option for analysing the selected image to obtain a preliminary classification. The user interface also provides a menu for selecting at least one user distinct from the authenticated user. Upon selection of the distinct user, the authenticated user is allowed to submit the preliminary classified image to the selected distinct user for re-classification dynamically. Finally, the user interface provides the user with an option for viewing the re-classified image real time. The user interface briefly described herein shall be described herein in detail.
The user interface is enabled to be used on devices selected from the list comprising but not limited to a desktop, a laptop, a scanner, a handled device such as a smart phone, a tablet, a fablet or a combination thereof. The user interface provides a menu for selecting a primary operation to be performed. In one example, the primary operation includes but is not limited to user authentication, a primary menu for selecting images or uploading an image, selecting a second user distinct from the first user for cross-examination/ re-classification, manually entering classification result or a combination thereof. The system described herein above briefly shall be described in detail.
In one specific example of the invention, the method and system, as described herein is used to differentiate between healthy red blood cells and malarial parasite infected red blood cells.
FIGS 3a – 3n generally show representative screen shots of the user interface, according to an embodiment of the invention. FIG. 3a shows the initial screen provided to a user. The user, upon authentication (not shown) is presented with options to select the test. FIG. 3b and 3c shows the selection of a scanned image of blood smear for testing malaria by a first user. Once the image is selected, the image is subjected to a preliminary classification as shown in FIG. 3d. The preliminary classification through automated process is achieved using a classification algorithm and takes about 10seconds – 50seconds. After the preliminary classification is done by the analyser, the results of the preliminary classification are displayed to the user as shown in FIG. 3e. The cells are classified by the classification algorithm as healthy, infected or unknown and are placed in different folders.
The first user now submits the preliminary classification for review to a second user. FIG. 3f shows a list of second users available to the first user for cross-examination of the preliminary classification. The first user can select one or more second users as shown in FIG. 3g. After sending the images for cross-examination, the second user receives the image as new message as shown in FIG. 3h and FIG. 3i. The second user, selects a folder for cross-validation. The second user then re-classifies the images in the unknown folder as shown in FIG. 3j and FIG. 3k. The final classification report is generated as shown in FIG. 3l and FIG. 3m. The final classification is sent to the first user as shown in FIG. 3n.
The advantage of the current invention is that, it provides for classification of cells on a real-time by basis. The classification system can be accessed at any remote location via a network. The total time taken to generate preliminary classification is about 10seconds-50seconds. Also, the method provides the ease for classification of a biological sample by a person not skilled in the art. Finally, this method enables upgradation of classification algorithm on a real-time basis.
The foregoing description of the invention has been set merely to illustrate the invention and is not intended to be limiting. Since modifications of the disclosed embodiments incorporating the spirit and substance of the invention may occur to person skilled in the art, the invention should be construed to include everything within the scope of the appended claims and equivalents thereof.
,CLAIMS:We claim:
1) A method for real-time classification of a biological sample, the method comprising:
selecting an image of the biological sample;
obtaining a preliminary classification of the image;
analysing the preliminary classification to determine accuracy of the classification; and
iteratively analyzing the preliminary classification to obtain a final classification.
2) The method of claim 1, wherein the image selected is one obtained by scanning the biological sample or from a database of pre-scanned images.
3) The method of claim 1, wherein the biological sample is prepared from a list comprising of a cell, a tissue, a smear of cells, a thin blood film, a thick blood film, a Pap smear, a sputum, a biological fluid, a cerebrospinal fluid, a fluid extracted from inflammatory site, a fluid extracted from tumor site or a fluid containing pathogens.
4) The method of claim 1, wherein the images are classified as belonging to a particular cell class or a cell type.
5) The method of claim 1, wherein the preliminary classification is a manual classification or an automated classification.
6) The method of claim 5, wherein the manual preliminary classification is done by a first user.
7) The method of claim 5, wherein the automated preliminary classification is obtained via a classification algorithm.
8) The method of claim 1, wherein the time taken for preliminary classification is in the range of about 10 seconds to about 50 seconds.
9) The method of claim 1, wherein the iterative analysis is a cross-examination of the preliminary classification by a second user distinct from the first user.
10) The method of claim 1, wherein the iterative analysis is a cross-examination of the preliminary classification by a plurality of users distinct from the first user and the second user.
11) The method of claim 10, wherein the cross-examination of the preliminary classification involves:
assigning scores to the classification of plurality of users;
obtaining a weighted average score; and
ranking the weighted average score to obtain a final classification of the sample.
12) A system for real-time classification of a biological sample, the system comprising:
a plurality of source devices;
a storage unit operably coupled to each of the source device;
an analyzer connected to the storage unit; and
a display unit coupled to the analyzer.
13) The system of claim 12, wherein the source devices are selected from a list comprising of a whole slide scanner, a scanning microscope, an imaging flow cytometer, an optofluidic microscope, a slide profiler or other known devices for cytological imaging.
14) The system of claim 12, wherein the storage unit comprises of a resident server, a remote server, a web server, a virtual server, a cloud or a combination thereof.
15) The system of claim 14, wherein the source devices are connected to the server through a local area network (LAN), a wide area network (WAN), a wireless network or a satellite network.
16) The system of claim 14, wherein the source devices are configured for two way communication with the server.
17) The system of claim 12, wherein the analyser is situated at the server, the source or at both the server and the source.
18) The system of claim 12, wherein the analyser comprises of a classification engine and a validation engine.
19) An interactive user interface for real-time classification of a biological sample, the interface comprising of:
a user authentication;
a menu for selecting an image;
analysing the selected image to obtain a preliminary classification;
a menu for selecting at least one user distinct from the authenticated user;
submitting the preliminary classified image to the selected distinct user for re-classification dynamically; and
means for viewing the re-classified image real time.
20) The user interface of claim 19, wherein the user interface is enabled on at least one device, wherein the device is at least one selected from a list comprising of a handheld device, a desktop device, a scanner and a laptop device.