A system and a method for monitoring peritoneal dialysis (PD) effluent to determine peritonitis condition using a machine learning model

Abstract

The present invention provides a system and a method for monitoring Peritoneal dialysis (PD) effluent to determine peritonitis condition using a machine learning model. The system comprises a user input device configured to upload at least one image of the peritoneal dialysis (PD) effluent bag and an image processing server having a processor configured to receive at least one image from the input device, process the received image of the PD effluent bag, by a convolutional neural network (CNN) model and classify the processed received image in one or more category of infection levels. The system further comprises an output device coupled to the image processing server and is configured to output the category of an infection level based on the processing of the image from the input device. Further, the processor is configured to create a training dataset from a plurality of images of the peritoneal dialysis (PD) effluent bag having an effluent density with varying colour gradient, train the convolutional neural network (CNN) model, with the plurality of the images of the PD effluent bag and classify the images of the PD effluent, with one or more category of the infection level.

Description

A SYSTEM AND A METHOD FOR MONITORING PERITONEAL DIALYSIS (PD) EFFLUENT TO DETERMINE PERITONITIS CONDITION USING A MACHINE LEARNING MODELTECHNICAL FIELD

[0001] Present disclosure relates to utilizing a machine learning model for determining peritonitis condition. More particularly, the present disclosure provides a system and a method for monitoring peritoneal dialysis (PD) effluent to determine peritonitis condition using a machine learning (ML) model.BACKGROUND

[0002] It is generally known that chronic kidney diseases are treated by dialysis, particularly peritoneal dialysis (PD). Peritoneal dialysis is a way to remove waste products from the blood. In peritoneal dialysis, a dialysate fluid flows through a tube into part of the stomach area i.e., abdomen. The inner lining of the abdomen, known as the peritoneum, acts as a filter and removes wastes from blood. After a set amount of time, the fluid with the filtered waste, i.e., PD effluent flows out of the abdomen and is thrown away.

[0003] However, said treatment procedure often results in complications such as peritonitis, which is commonly occurring complication and can lead to put the peritoneal dialysis on hold or discontinued unless such complications are obviated. One common indicator which reflects peritonitis condition is colour of the PD effluent. As generally known, peritoneal effluent contains electrolytes and uremic toxins and if it is persistent for a while, the drained PD effluent would have a colour similar to yellow. Further, effluent that has been dwelling with toxins for a long time would become cloudy and often, late diagnosis of peritonitis can lead to death.

[0004] One common method of analysis PD effluent to check the peritonitis condition, is to send the PD effluent bag which collects the PD fluid to a laboratory where the lab specialist will use the state-of-the-art method to determine the level ofinfection based on effluent bag fluid. However, this method has a disadvantage of longer turn-around time, since the time taken for PD effluent bag collection, sending it to the laboratory, analysing of PD effluent fluid and providing the results to the patient may take days depending upon the location of the patient. Further, as this is a continuous process, following existing procedure for each PD effluent bag may result in failure to detect early infection warning and may result in increase in infection level due to lack of timely intervention.

[0005] Therefore, there exists a need for a technology which can detect early signs of complications during PD and reduce manual invention in handling PD effluent bag and to overcome deficiencies of conventional peritonitis detection methods.

[0006] Applicant has identified many technical challenges and difficulties associated with current solutions and through applied effort, ingenuity, and innovation, the applicant has provided a solution to the above-mentioned drawbacks.SUMMARY OF THE INVENTION

[0007] In general, embodiments of the present disclosure herein provide a system and a method for monitoring peritoneal dialysis (PD) effluent to determine peritonitis condition using a machine learning (ML) model. Other implementations will be, or will become, apparent to one with skill in the art upon examination of the following figures and detailed description. It is intended that all such additional implementations be included within this description be within the scope of the disclosure.

[0008] According to one embodiment, the present invention provides a system for monitoring Peritoneal dialysis (PD) effluent to determine peritonitis condition using a machine learning model. The system comprises a user input device configured to upload at least one image of the peritoneal dialysis (PD) effluent bag. The system further comprises a processor, a memory storing machine readable instructions, when executed cause the processor to receive at least one image from the input device, process the received image of the PD effluent bag, by a convolutional neural network(CNN) model and classify the processed received image in one or more category of infection levels. The system also comprises an output device which is coupled to the image processing server and is configured to output the category of an infection level based on the processing of the image from the input device.

[0009] In another aspect of the present invention, the processor is configured to create a training dataset from a plurality of images of the peritoneal dialysis (PD) effluent bag having varying colour gradient, train the convolutional neural network (CNN) model, with the plurality of the images of the PD effluent bag and classify, each of the images of the PD effluent, with one or more category wherein each category indicates the infection level. In an embodiment, the categories of the infection level are any one of normal or peritoneal condition.

[0010] In another aspect of the present invention, the input device is a mobile device and is configured to capture the image the PD effluent bag using the device camera. Further, the output device is coupled to the user input device and is configured to notify the infection level on the user input device. Further, the output device is further configured to send one or more recommendations to the input device, indicating the need for clinical review.

[0011] In yet another aspect, the image processing server is either hosted on a cloud network or on a local network.

[0012] In an additional embodiment, the user input device is further configured to upload a biomarker test results in addition to the PD effluent image.

[0013] In an aspect of the present invention, the image processing server is configured to validate the output of the convolutional neural network (CNN) model indicating the category of the infection level.

[0014] In another embodiment, the present invention provides a method for monitoring Peritoneal dialysis (PD) effluent to determine peritonitis condition using a machine learning model, comprising uploading, by a user input device, at least one image of the peritoneal dialysis (PD) effluent bag; receiving, by an image processing server, at least one image from the input device; sending, by the image processing server, at least one image from the user input device to a convolutional neural network (CNN) model; processing, by the CNN model, the received image of the PD effluent bag; classifying, by the CNN model, the processed received image in one or more category of infection levels; and outputting, by the image processing server, the category an infection level on an output device.

[0015] In an embodiment, the method comprises creating, by image processing server, a training dataset from a plurality of images of the peritoneal dialysis (PD) effluent bag having varying colour gradient; training, the convolutional neural network (CNN) model, with the plurality of the images of the PD effluent; and classifying, each of the images of the PD effluent, with one or more category wherein each category indicates the infection level.

[0016] In yet another embodiment, the method comprises notifying the infection level on the user input device. The method further comprises uploading, by the user input device, a biomarker test results in addition to the PD effluent bag image.

[0017] In yet another embodiment, the method comprises validating, by the image processing server, the output of the convolutional neural network (CNN) model indicating the category of the infection level. Further, the method comprises sending one or more recommendations to the input device, said recommendations indicating a need for clinical review.

[0018] In yet another embodiment, the present invention provides a method for training and deploying a convolutional neural network (CNN) model for predicting the infection level of Peritoneal dialysis (PD) effluent, comprising: creating, by an imageprocessing server, a training dataset from a plurality of images of the peritoneal dialysis (PD) effluent bag having effluent density with varying colour gradient; training, the convolutional neural network (CNN) model, with the plurality of the image of the PD effluent; and classifying, each of the images of the PD effluent bag, with one or more category wherein each category indicates the infection level.

[0019] In an embodiment, the convolutional neural network (CNN) model is any one of VGG16, Resnet 50 or EfficientNet.

[0020] In an embodiment, the method comprises pre-processing the plurality of images of the peritoneal dialysis (PD) effluent to resize all the images to a constant size of 400x400 pixels.

[0021] In another embodiment, the method comprises utilizing 60% of the images for training the CNN model, 15% of the images for validation purpose and 25% of the images for testing purpose.

[0022] In an embodiment, the method comprises performing data augmentation, wherein new image data is generated from existing image data by applying transformations like horizontal / vertical flip, zoom, rotation and translation.

[0023] In another embodiment, the CNN model comprises one or more layers comprising a convolutional layer for defining grids which are used that move over the image and each grid looks for a specific pattern like a line, curve or shapes, a pooling layer for extracting the most significant features by applying aggregation operations and fully connected layer for using the pooling layer output to classify the images. The method comprises, in one embodiment, deploying the convolutional neural network (CNN) model on a cloud network. In another embodiment, the convolutional neural network (CNN) model is deployed on a local network.

[0024] The above summary is provided merely for the purpose of summarizing some example embodiments to provide a basic understanding of some aspects of the present disclosure. Accordingly, it will be appreciated that the above-described embodiments are merely examples and should not be construed to narrow the scope or spirit of the present disclosure in any way.

[0025] It will be appreciated that the scope of the present disclosure encompasses many potential embodiments in addition to those here summarized, some of which will be further described below. Other features, aspects, and advantages of the subject will become apparent from the description, the drawings, and the claims.BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings constitute a part of the description and are used to provide further understanding of the present disclosure. Such accompanying drawings illustrate the embodiments of the present disclosure which are used to describe the principles of the present disclosure.

[0027] Fig. 1 illustrates an exemplary way of examining the peritoneal dialysis (PD) effluent, in accordance with the prior art;

[0028] Fig. 2 illustrates an exemplary system for determining peritonitis condition using a machine learning (ML) model, in accordance with an embodiment of the present disclosure;

[0029] Fig. 3 illustrates an exemplary images of PD effluent bags which are used for training of a Machine Learning (ML) model, in accordance with an embodiment of the present disclosure;

[0030] Fig. 4 illustrates an exemplary system for training the ML model, in accordance with an embodiment of the present disclosure;

[0031] Fig. 5 illustrates block diagram of training of ML model, in accordance with an embodiment of the present disclosure;

[0032] Fig. 6 illustrates the training and validation loss percentage while training the model, in accordance with an embodiment of the present disclosure;

[0033] Fig. 7 illustrates an exemplary user device and interface, in accordance with one embodiment of the present disclosure;

[0034] Fig. 8 illustrates a flow-chart for determining peritonitis condition using the input device, in accordance with one embodiment of the present disclosure;

[0035] Fig. 9 illustrates a flow-chart for training of ML model, in accordance with one embodiment of the present disclosure; and

[0036] Fig. 10 illustrates general architecture of the system for implementation, in accordance with an embodiment of the present invention.

[0037] The embodiments are illustrated by way of example and not by way of limitation in the figures of the accompanying drawings in which like references indicate similar elements. It should be noted that references to “an” or “one” embodiment in this disclosure are not necessarily to the same embodiment and they mean at least one.DETAILED DESCRIPTION OF THE INVENTION

[0038] Some embodiments of the present disclosure now will be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all, embodiments of the disclosure are shown. Indeed, embodiments of the disclosure may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein, rather, these embodiments are provided so that thisdisclosure will satisfy applicable legal requirements. Like numbers refer to like elements throughout.

[0039] As used herein, the term “comprising” means including but not limited to and should be interpreted in the manner it is typically used in the patent context. Use of broader terms such as comprises, includes, and having should be understood to provide support for narrower terms such as consisting of, consisting essentially of, and comprised substantially of.

[0040] The phrases “in one embodiment,” “according to one embodiment,” “in some embodiments,” and the like generally mean that the particular feature, structure, or characteristic following the phrase may be included in at least one embodiment of the present disclosure, and may be included in more than one embodiment of the present disclosure (importantly, such phrases do not necessarily refer to the same embodiment).

[0041] The word “example” or “exemplary” is used herein to mean “serving as an example, instance, or illustration.” Any implementation described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other implementations .

[0042] As known, peritoneal dialysis (PD) effluent’s density along with colour gradient is an indicator of infection level reflecting on abdomen infection. Fig. 1 illustrates an exemplary way of examining the peritoneal dialysis (PD) effluent, in accordance with the prior art. Commonly, the effluent is collected in an effluent bag and is sent for medical examination to a laboratory, where the necessary medical analysis is carried to assess the infection level and subsequently, clinical review is carried out based on medical reports. Often, the clinical review may mandate conducting biomarker assessment to supplement the medical examination of effluent, to ascertain the reasons for the infection level. This complete assessment procedures involves manual invention, transportation of effluent bags and subsequently, thereports are provided to the user via messages or physical reports, which is a timeconsuming procedure. This manual procedure may result in delayed detection of infection level, and often results in increase in infection level before proper medical invention is implemented.

[0043] Therefore, there is a need for an efficient methodology and system, which can be efficient in detecting early signs of complications during PD and reduce manual invention in handling PD effluent bag and to overcome deficiencies of conventional peritonitis detection methods.

[0044] Fig. 2 illustrates an exemplary system for determining peritonitis condition using a machine learning (ML) model, in accordance with an embodiment of the present disclosure. The system comprises an image capture device 203, an image processing module 204, a database 205 and an output portal 206. The image capture device 203 is configured to capture one or more images of an effluent bag 201, which contains peritoneal dialysis (PD) fluid. The image capture device 203 may be implemented as a mobile device, sometimes referred to as a smart phone. In another embodiment, the image capture device 203 may be implemented as, any type of wireless device, mobile station, or portable computing device with a self-contained power source (e.g., battery) such as a laptop computer, ultra-laptop computer, personal digital assistant (PDA), cellular telephone, combination cellular telephone / PDA, mobile unit, subscriber station, user terminal, portable computer, handheld computer, palmtop computer, wearable computer, media player, pager, messaging device, data communication device, and so forth.

[0045] In an embodiment, the effluent bag is a transparent bag which stores the PD fluid without having exposure to external environment. The transparent nature of PD effluent bag enables clear view of colour gradient of the PD fluid. The one or more images of effluent bag is stored in a memory of the image capturing device as an image file. In an embodiment, the image file is any one of JPEG, PNG.

[0046] The image capture device 203 is coupled to the image processing module 204, wherein the image processing module 204 further comprises a processor 204a, a memory 204b and a machine learning model 204c. In an example, the processor 204a may be a single processing unit or a number of units, all of which could include multiple computing units. The processor may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, logical processors, virtual processors, state machines, logic circuitries, and / or any devices that manipulate signals based on operational instructions. Among other capabilities, the processor is configured to fetch and execute computer-readable instructions and data stored in the memory 204b. In an embodiment, the memory 204b may include any non-transitory computer-readable medium known in the art including, for example, volatile memory, such as static random-access memory (SRAM) and dynamic random-access memory (DRAM), and / or non-volatile memory, such as read-only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes.

[0047] A user 200 is illustrated in Fig. 2 with peritonitis condition and the user 200 (may also be referred to as ‘a patient’) collects the Peritoneal dialysis (PD) effluent in a transparent bag 201. The user 200, via the user device 203, captures the image of the effluent transparent bag and the image of the effluent bag is stored in a memory of the user device 203. In an embodiment, the image of the effluent bag is an actual colour image of the effluent bag. In another embodiment, the user 200 may also upload a biomarker results 202 in the user device 203 and is stored in the user device’ s memory.

[0048] In one embodiment, the user device 203 is configured to download and install an application on the user device 203, which provides one or more provisions to upload one or more images of PD effluent. In another embodiment, a user interface of the user device 203 may be provided with a provision to directly capture and upload the one or more images of PD effluent bag. In an embodiment, any image that isblurred is validated by the application and the user would be prompted to re-upload a valid effluent bag image.

[0049] One or more images of the PD effluent bag 201 is uploaded to the image processing module 204. The processor 204a receives the uploaded images of PD effluent bag 201, and process the image to remove noise. In an embodiment, the processing of the images of PD effluent may include augmentation of the uploaded images. The processed image is then provided to the machine learning (ML) model 204c for further processing. In one embodiment, the machine learning (ML) model is a convolutional neural network (CNN) model.

[0050] The machine learning (ML) model 204c receives at least one image of the PD effluent and processes the received image of the PD effluent bag to classify the processed received image in one or more category of infection levels. In one embodiment, the inflection level may include any one of normal condition or peritonitis condition. In another embodiment, the inflection level may include any one of mild condition, moderate condition or severe condition. The image processing module 204 further comprises a memory 204b which is configured to store one or more user profiles and classification of PD effluent bag with a time stamp.

[0051] The image processing module 204 is further coupled to the output module 206, the output module 206 is configured to provide a report to the user through the interface of the user device. In an embodiment, the output of the ML model, including classification of PD effluent bag and its summary, is provided as a generated report to the user 200. In an embodiment, the report is provided in the form of a summary on the user interface. In another embodiment, the report is provided in the form of a downloadable file through the user interface.

[0052] In a further embodiment, the summary on the user interface may include a recommendation for uploading a biomarker test report, for further analysis. Uponuploading the biomarker test report on the user device, the same is provided to the ML model for further analysis and update the summary on the user interface.

[0053] The user device 203 may provide one or more options for providing the user the access to the application and enables uploading of images. In one embodiment, a windows application may be provided on the user device which uses the python library (tkinter) to allow the user to login. The user interface uses Python REST API to transfer the effluent image and / or biomarker data to a Spring boot server. The windows application may have a clinician mode, which uses the Python REST API to show the notification in case the model predicted the effluent image as peritonitis.

[0054] The image processing module 204 is further coupled the database 205. The database 205 is configured to store one or more images of PD effluent bag of varying colour gradients. As an example, the database 205 may be implemented with integrated hardware and software. The hardware may include a hardware disk controller with programmable search capabilities or a software system running on general-purpose hardware. Examples of databases are but are not limited to, in-memory databases, cloud databases, distributed databases, embedded databases, and the like. The database amongst other things, serves as a repository for storing data processed, received, and generated by one or more of the processors, and the modules / engines / units.

[0055] In one embodiment, the processor 204a is coupled to the database 205 and the processor 204a is configured to fetch one or more images from the database 205, referred to as “training data set” and train the machine learning model 204c based on the training data set. Aspects of model training will be discussed henceforth with reference to Fig. 3 and Fig. 4.

[0056] Fig. 3 illustrates an exemplary images of PD effluent bags which are used for training of a Machine Learning (ML) model, in accordance with an embodiment of the present disclosure. In an embodiment, the PD effluent bag are transparent bag which stores the PD effluent fluid from the user and is captured by an image capturedevice. Each of the PD effluent bag containing the fluid may have different effluent density along with colour gradient, based on the infection level. The effluent bag with clear or pale yellow fluid may indicate normal condition, but the effluent bag with cloudy colour gradient may indicate the infection level. For example, reddish peritoneal fluid may indicate the presence of blood, most often due to trauma and cloudy peritoneal fluid may indicate the presence of microbes and / or white blood cells (WBCs), pointing to an infection.

[0057] In an embodiment, one or more images of the effluent bags containing fluid are stored in the database 205 and is used by the processor 204a in training the ML model 204c.

[0058] Fig. 4 illustrates an exemplary system for training the ML model, in accordance with an embodiment of the present disclosure. In one embodiment, the model uses Convolutional Neural Network (CNN), a type of deep learning algorithm which consists of layers: convolutional layer, pooling layer and fully connected layer and is used for image classification, object identification etc. CNNs are distinguished from other machine learning algorithms by their ability to autonomously extract features efficiently at a large scale, allowing us to bypass manual feature engineering.

[0059] A training system is illustrated which comprises a database 402, a processor 403 and a Machine Learning Model 404. One or more images of the effluent bags containing fluid is stored in the database 402 and is used by the processor 403 in training a Machine Learning (ML) model 404. The ML model 404, in an embodiment, comprises a Machine Learning Algorithm module 404a, a test and tuning module 404b and an output module 404c.

[0060] The machine learning algorithm module 404a may comprises one or more computational models that allows to understand patterns, colours, colour gradients and make judgments based on data. One or more machine learning algorithms form the foundation of artificial intelligence and are used image recognition. Further, themachine learning algorithm module 404a is coupled to the test and tuning module 404b. Further, the machine learning model 404 comprises an output module 404c, the output module 404c is coupled to the test and tuning module 404b.

[0061] The processor 403 is configured to receive plurality of the images of PD effluent bags, which forms the training dataset for the machine learning and performs the image augmentation. As part of pre-processing, the processor 403 resizes the images to a constant size 400x400, and the images are then split into training, validation and test set. In one embodiment, the resized images are split into training, validation and test set with a split ratio of 60%-15%-25%.

[0062] Fig. 5 illustrates a block diagram of training of the ML model, in accordance with an embodiment of the present disclosure. In one embodiment, the processor 403, through image augmentation, generate a new image data from existing image data 401 by applying transformations like horizontal / vertical flip, zoom, rotation and translation. In an embodiment, the data augmentation is applied on training and validation set and the test set is left unchanged.

[0063] The machine learning algorithm module 404a is coupled to the training and tuning module 404b. In an embodiment, the training and tuning module 404b is trained with the dataset of labelled effluent bag images, and comprises at least one convolutional layer, at least one pooling layer and a fully connected layer. In the convolutional layer, small grids (called filters / kemels) are used that move over the image. Each grid looks for a specific pattern like a line, curve or shapes. In an embodiment, one filter may be good at finding lines while the other at finding curve or shapes. Using different filters, the neural network will have a good assessment of all the patterns that make up the image.

[0064] Further, the pooling layer then pulls the most significant features by applying aggregation operations. The Fully connected layer uses the pooling layer output to classify the images to classes: Peritonitis or Normal. In an embodiment, afully connected layer uses SoftMax activation function to classify the image as Peritonitis or Normal.

[0065] In an embodiment, one or more CNN models are used and may be any one of EfficientNetB7, Resnet50 and VGG16. In a preferred embodiment, EfficientNetB7 version of the model is used. The CNN model is trained with dataset including effluent images where the model learns to classify the low-level or generic features such as colours, comers, shapes, edges etc.

[0066] In another embodiment, a spring boot is provided which is a web application that includes an embedded web server. It takes the effluent image as input from the mobile and web application and store it in a database. A Sprint batch which is a lightweight batch framework has the ability to efficiently process large amount of data and the spring batch server continuously polls the database and when a new record is added in the database, it transfers this record to a Flask application. The model is pretrained with effluent images and the pretrained model is integrated with Flask Application using Keras Python library. In one embodiment, Keras Python Library is used to augment the data (i.e., rotate, scale, flip) effluent image samples. In an embodiment, one or more pretrained models may be used which may be one of VGG16, Resnet 50 and other EfficientNet with transfer learning. Transfer learning enables to train the model with various pre-trained weights to obtain the best performance.

[0067] The table below illustrates exemplary model and their total number of parameters in millions (M).*Parameters in general are weights that are learnt (learnable elements) during training.

[0068] The machine learning module 404 further comprises an output module 404c which is configured to generate a report for the user’s review on the user interface. In an embodiment, the output module 404c is also coupled to a portal of clinic, wherein the reports of the model are presented to the doctor or practitioner handling the concerned patient for clinic review and recommendations.

[0069] Fig. 6 illustrates the training and validation loss percentage while training the model, in accordance with an embodiment of the present disclosure. In an embodiment, Binary Cross Entropy is selected as the loss function (aids the model in penalizing the inaccurate prediction), Adam as the optimizer (updates the model parameters based on the loss function output) and a fixed learning rate, and then train the model for 100 epochs. Once the training is done, the model is evaluated on the test set and a confusion matrix is calculated.

[0070] The Confusion matrix of various model is provided below:

[0071] Fig. 7 illustrates an exemplary user device 700 and interface, in accordance with one embodiment of the present disclosure. The user device 700 is configured to download and install an application which provides the interface to the user to perform one or more functions associated with examination of PD effluent bags. In anembodiment, the user interface provides a block 701 which provides the provisions for the user to click and upload a picture of PD effluent bag. In a further embodiment, the icon 702 may enable the user to select and upload an image from the gallery of the user device.

[0072] Once one or more images are submitted, the portion of the user interface 703 may provide summary of the assessment after the machine learning model processes the image and classify the image in one of the categories. Further, the portion 704 of the user interface may display one or more recommendations to the user based on assessment. In one embodiment, the recommendation may include need for immediate doctor visit, if it is determined that the infection level is high. In another embodiment, the recommendations may include any one of uploading of biomarker test results, clinic review recommendations provided by a clinic practitioner once the classification results are provided for clinical review by the practitioner.

[0073] In a further embodiment, the portion 705 provides an option for reassessment of the classification, if the user is not satisfied by the output of the machine learning model.

[0074] Fig. 8 illustrates a flow-chart for determining peritonitis condition using the input device, in accordance with one embodiment of the present disclosure. In step 801, one or more images of the peritoneal dialysis (pd) effluent bag is uploaded on the user device. In an embodiment, the user device may upload and install an application which provides one or more options to upload the images of the effluent bag.

[0075] In Step 802, the processor of the image processing module 204 receives the image from the user device. The processor 204a receives the uploaded images of PD effluent, and process the image to remove noise. In an embodiment, the processing of the images of PD effluent may include augmentation of the uploaded images.

[0076] In Step 803, the processed image is provided to the machine learning (ML) model 204c for further processing.

[0077] In Step 804, the received images of the PD effluent is processed by the machine learning model. In one example, the machine learning model is a convolutional neural network (CNN) model. In one embodiment, the machine learning model access the patterns, colours, colour gradients and make judgments based on data.

[0078] In Step 805, the machine learning model classify the processed image in one or more category of infection level by the machine learning model. The machine learning (ML) model 204c classifies the processed received image in one or more category of infection levels. In one embodiment, the inflection level may include any one of normal condition or peritonitis condition. In another embodiment, the inflection level may include any one of mild condition, moderate condition or severe condition. In a further embodiment, the image processing module 204 is further coupled to the output module 206, the output module is configured to provide a report to the user through the interface of the user device.\

[0079] In Step 806, the user device is provided with indication of the infection level based on processing of image by the machine learning model. In one embodiment, once one or more images are submitted, the portion 703 of the user interface may provide summary of the assessment after the machine learning model processes the image and classify the image in one of the categories. Further, the portion 704 of the user interface may display one or more recommendations to the user based on assessment. In one embodiment, the recommendation may include need for immediate doctor visit, if it is determined that the infection level is high. In another embodiment, the recommendations may include any one of uploading of biomarker test results, clinic review recommendations provided by a practitioner once the classification results are provided for clinic review to the practitioner.

[0080] Fig. 9 illustrates a flow-chart for training of ML model, in accordance with one embodiment of the present disclosure.

[0081] In Step 901, a training data set 401 is created from a plurality of images of the peritoneal dialysis (pd) effluent bag having varying colour gradient. In an embodiment, the database 402 configured to store one or more images of PD effluent bag of varying colour gradients. As an example, the database 402 may be implemented with integrated hardware and software. In another embodiment, one or mora machine Learning model is trained. In an embodiment, the machine learning model is a convolutional neural network (CNN) model. In a further embodiment, the processor 403 is configured to receive and process the input images of plurality of PD effluents bags and perform the image augmentation. As part of pre-processing, the processor 403 resizes the images to a constant size 400x400, and the images are then split into training, validation and test set. In one embodiment, the processor, through image augmentation, generate a new image data from existing image data by applying transformations like horizontal / vertical flip, zoom, rotation and translation. The data augmentation is applied on training and validation set and the test set is left unchanged.

[0082] In step 902, the Machine Learning Model is trained with the plurality of the images of the PD effluent. In an embodiment, the machine learning model may comprise one or more computational models that allows to understand patterns, colours, colour gradients and make judgments based on data.

[0083] In an embodiment, the machine learning model is trained with dataset of labelled effluent bag images, and comprises at least one Convolutional layer, Pooling layer and a fully connected layer. In a convolutional layer, small grids (called filters / kernels) are used that move over the image. Each grid looks for a specific pattern like a line, curve or shapes. In an embodiment, one filter might be good at finding lines while the other at finding curve or shapes. Using different filters, the CNN will have a good assessment of all the patterns that make up the image. Further, the pooling layer then pulls the most significant features by applying aggregation operations.

[0084] In Step 903, the image of the PD effluent bag is classified with one or more category of infection level. In an embodiment, the CNN comprises a Fully connected layer which uses the pooling layer output to classify the images to classes: Peritonitis or Normal. In an embodiment, a fully connected layer uses SoftMax activation to classify the image as Peritonitis or Normal.

[0085] Fig. 10 illustrates general architecture of the system 1000 for implementation, in accordance with an embodiment of the present invention.

[0086] In an example, the module(s), engine(s), and / or unit(s) 1002 may include a program, a subroutine, a portion of a program, a software component or a hardware component capable of performing a stated task or function. As used herein, the module(s), engine(s), and / or unit(s) may be implemented on a hardware component such as a server independently of other modules, or a module can exist with other modules on the same server, or within the same program. The module (s), engine(s), and / or unit(s) 1002 may be implemented on a hardware component such as processor one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, and / or any devices that manipulate signals based on operational instructions. The module (s), engine(s), and / or unit(s) 1002 when executed by the processor(s) 1002 may be configured to perform any of the described functionalities. According to an embodiment, the module 1002 includes one or more modules discussed above. In an alternate embodiment, the functions of the aforesaid modules may be performed by the processor(s) 1001.

[0087] The modules / engines / units 1002 may be implemented with an Al module that may include a plurality of neural network layers. Examples of neural networks include, but are not limited to, a convolutional neural network (CNN), a deep neural network (DNN), a recurrent neural network (RNN), a Restricted Boltzmann Machine (RBM). The learning technique is a method for training a predetermined target deviceusing a plurality of learning data to cause, allow, or control the target device to decide or prediction. Examples of the learning techniques include, but are not limited to, a supervised learning, unsupervised learning, a semi-supervised learning, or reinforcement learning. At least one of a plurality of CNN, DNN, RNN, RMB models and the like may be implemented to thereby achieve execution of the present subject matter’s mechanism through an Al model. A function associated with the Al model may be performed through the non-volatile memory, the volatile memory, and the processor. The processor may include one or a plurality of processors. At this time, one or a plurality of processors may be a general-purpose processor, such as a central processing unit (CPU), an application processor (AP), or the like, a graphics-only processing unit such as a graphics processing unit (GPU), a visual processing unit (VPU), and / or an Al-dedicated processor such as a neural processing unit (NPU). The one or a plurality of processors control the processing of the input data in accordance with a predefined operating rule or the artificial intelligence (Al) model stored in the non-volatile memory and the volatile memory. The predefined operating rule or artificial intelligence model is provided through training or learning.

[0088] As an example, the display unit 1004 includes a computer monitor, a touch screen, an output device capable of displaying the graphics, and the like. The display unit 1004 is configured to display visual output in desktops, laptops, and workstations.

[0089] As a further example, the network interface 1006 is configured to provide and establish communication with any electronic device via a public network, private network, or any wireless communication technology.

[0090] The present invention presents various advantages, which enables early infection indication for the patient as well as, for the clinic and provides constant monitoring of the peritoneum status. Some of the advantages of the present invention may include sending reminders to patients to perform periodic check, generate alerts to clinic thereby helping clinician to make timely decisions and instruct patient onantibiotics medication, on time treatment that prevents removal of catheter, home based check to avoid visits to the hospital.

[0091] The figures of the disclosure are provided to illustrate some examples of the invention described. The figures are not to limit the scope of the depicted embodiments or the appended claims. Aspects of the disclosure are described herein with reference to the invention to example embodiments for illustration. It should be understood that specific details, relationships, and method are set forth to provide a full understanding of the example embodiments. One of ordinary skill in the art recognize the example embodiments can be practiced without one or more specific details and / or with other methods.

[0092] Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of various system components in the embodiments described above should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0093] Aspects of the present disclosure may be implemented as computer program products that comprise articles of manufacture. Such computer program products may include one or more software components including, for example, applications, software objects, methods, data structure, and / or the like. In some embodiments, a software component may be stored on one or more non-transitory computer-readable media, which computer program product may comprise the computer-readable media with software component, comprising computer executable instructions, included thereon. The various control and operational systems described herein may incorporate one or more of such computer program products and / orsoftware components for causing the various conveyors and components thereof to operate in accordance with the functionalities described herein.

[0094] A software component may be coded in any of a variety of programming languages. An illustrative programming language may be a lower- level programming language such as an assembly language associated with a particular hardware architecture and / or operating system platform / system. Other example of programming languages included, but are not limited to, a macro language, a shell or command language, a job control language, a script language, a database query, or search language, and / or report writing language. In one or more example embodiments, a software component comprising instructions in one of the foregoing examples of programming languages may be executed directly by an operating system or other software component without having to be first transformed into another form. A software component may be stored as a file or other data storage methods. Software components of a similar type or functionally related may be stored together such as, for example, in a particular directory, folder, or repository. Software components may be static (e.g., pre-established, or fixed) or dynamic (e.g., created or modified at the time of execution).

[0095] While this specification contains many specific implementation details, these should not be construed as limitations on the scope of any disclosures or of what may be claimed, but rather as descriptions of features specific to particular embodiments of particular disclosures. Certain features that are described herein in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable sub combination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a sub combination or variation of a sub combination.

[0096] Thus, particular embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In certain implementations, multitasking and parallel processing may be advantageous.

[0097] It is to be understood that the disclosure is not to be limited to the specific embodiments disclosed, and that modifications and other embodiments are intended to be included within the scope of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation, unless described otherwise.

[0098] A processor may include one or more general purpose processors (e.g., INTEL® or Advanced Micro Devices® (AMD) microprocessors) and / or one or more special purpose processors (e.g., digital signal processors or Xilinx® System On Chip (SOC) Field Programmable Gate Array (FPGA) processor), MIPS / ARM-class processor, a microprocessor, a digital signal processor, an application specific integrated circuit, a microcontroller, a state machine, or any type of programmable logic array.

[0099] A memory may include, but is no limited to, non-transitory machine- readable storage devices such as hard drives, magnetic tape, floppy diskettes, optical disks, Compact Disc Read-Only Memories (CD-ROMs), and magneto-optical disks, semiconductor memories, such as ROMs, Random Access Memories (RAMs), Programmable Read-Only Memories (PROMs), Erasable PROMs (EPROMs), Electrically Erasable PROMs (EEPROMs), flash memory, magnetic or optical cards, or other type of media / machine-readable medium suitable for storing electronic instructions.

[0100] The terms operator, person, employee, worker, labour, personnel, workforce, supervisor, and manager have been used interchangeably throughout the draft and corresponds to an individual working in an industry or directly or indirectly managing operations in the industrial environment.

[0101] The terms “or” and “and / or” as used herein are to be interpreted as inclusive or meaning any one or any combination. Therefore, “A, B or C” or “A, B and / or C” mean “any of the following: A; B; C; A and B; A and C; B and C; A, B and C.” An exception to this definition will occur only when a combination of elements, functions, steps or acts are in some way inherently mutually exclusive.

[0102] Any combination of the above features and functionalities may be used in accordance with one or more embodiments. In the foregoing specification, embodiments have been described with reference to numerous specific details that may vary from implementation to implementation. The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense. The sole and exclusive indicator of the scope of the invention, and what is intended by the applicants to be the scope of the invention, is the literal and equivalent scope of the set as claimed in claims that issue from this application, in the specific form in which such claims issue, including any subsequent correction.

Claims

We Claim:

1. A system for monitoring Peritoneal dialysis (PD) effluent to determine peritonitis condition using a machine learning model, comprising: a user input device configured to upload at least one image of a peritoneal dialysis (PD) effluent bag; an image processing server comprising: a processor; a memory storing machine readable instructions, when executed cause the processor to: receive at least one image from the user input device; process the received image of the PD effluent bag, by a convolutional neural network (CNN) model; classify the processed received image in one or more category of infection levels; and an output device coupled to the image processing server and is configured to output a category of an infection level based on the processing of the image from the input device.

2. The system of claim 1, wherein the processor is configured to: create a training dataset from a plurality of images of the peritoneal dialysis (PD) effluent bag having an effluent density with varying colour gradient; training the convolutional neural network (CNN) model, with the plurality of the images of the PD effluent bag; and classify, each of the images of the PD effluent, with one or more category wherein each category indicates the infection level.

3. The system of claim 1, wherein the categories of the infection level is any one of normal or peritoneal condition.

4. The system of claim 1, wherein the output device is further coupled to the user input device and is configured to notify the infection level on the user input device.

5. The system of claim 1, wherein the image processing server is hosted on a cloud network.

6. The system of claim 1 , wherein the image processing server is hosted on a local network.

7. The system of claim 1, wherein the user input device is further configured to upload a biomarker test results in addition to the PD effluent image.

8. The system of claim 1, wherein the image processing server is configured to validate the output of the convolutional neural network (CNN) model indicating the category of the infection level.

9. The system of claim 1, wherein the output device is further configured to send one or more recommendations to the input device, indicating a need for clinical review.

10. The system of claim 1, wherein the input device is a mobile device and is configured to capture the image the PD effluent bag using a device camera.

11. A method for monitoring Peritoneal dialysis (PD) effluent to determine peritonitis condition using a machine learning model, comprising: uploading, by a user input device, at least one image of the peritoneal dialysis (PD) effluent bag; receiving, by an image processing server, at least one image from the input device;sending, by the image processing server, at least one image from the user input device to a convolutional neural network (CNN) model; processing, by the CNN model, the received image of the PD effluent bag; classifying, by the CNN model, the processed received image in one or more category of infection levels; and outputting, by the image processing server, the category an infection level on an output device.

12. The method of claim 11, comprises: creating, by image processing server, a training dataset from a plurality of images of the peritoneal dialysis (PD) effluent bag having an effluent density with varying colour gradient; training, the convolutional neural network (CNN) model, with the plurality of the images of the PD effluent; and classifying, each of the images of the PD effluent, with one or more category wherein each category indicates the infection level.

13. The method of claim 11, wherein the categories of the infection level is any one of normal or peritoneal condition.

14. The method of claim 11, comprising: notifying the infection level on the user input device.

15. The method of claim 11, wherein the image processing server is hosted on a cloud network or a local network.

16. The method of claim 11, further comprising: uploading, by the user input device, a biomarker test results in addition to the PD effluent bag image.

17. The method of claim 11, further comprising: validating, by the image processing server, the output of the convolutional neural network (CNN) model indicating the category of the infection level.

18. The method of claim 11, further comprising: sending one or more recommendations to the input device, said recommendations indicating a need for clinical review.

19. A method for training and deploying a convolutional neural network (CNN) model for predicting the infection level of Peritoneal dialysis (PD) effluent, comprising: creating, by an image processing server, a training dataset from a plurality of images of the peritoneal dialysis (PD) effluent bag having an effluent density with varying colour gradient; training, the convolutional neural network (CNN) model, with the plurality of the image of the PD effluent; and classifying, each of the images of the PD effluent bag, with one or more category wherein each category indicates the infection level.

20. The method of claim 19, wherein the convolutional neural network (CNN) model is any one of VGG16, Resnet 50 or EfficientNet.

21. The method of claim 19, comprising: pre-processing the plurality of images of the peritoneal dialysis (PD) effluent to resize all the images to a constant size of 400x400 pixels.

22. The method of claim 19, comprising: utilizing 60% of the images for training the CNN model; utilizing 15% of the images for validation purpose; and utilizing 25% of the images for testing purpose.

23. The method of claim 19, comprising: performing data augmentation, wherein new image data is generated from existing image data by applying transformations like horizontal / vertical flip, zoom, rotation and translation.

24. The method of claim 19, wherein the CNN model comprises one or more layers comprising: a convolutional layer for defining grids which are used that move over the image and each grid looks for a specific pattern like a line, curve or shapes; a pooling layer for extracting the most significant features by applying aggregation operations; and fully connected layer for using the pooling layer output to classify the images.

25. The method of claim 19, comprising: deploying the convolutional neural network (CNN) model on a cloud network.

Images