<\/h6>\r\nIntroduction<\/h2>\r\n<\/h6>\r\nDiabetic Retinopathy (DR) is one of the leading cause for blindness, affecting over 93 million people across the world. DR is an eye disease associated with diabetes. Detection and grading DR at an early stage would help in preventing permanent vision loss. Automated detection and grading during the retinal screening process would help in providing a valuable second opinion. In this blog, we implement a simple transfer-learning based approach using a deep Convolutional Neural Network (CNN) to detect DR<\/strong>.\r\n<\/h6>\r\nPlease cite the following articles if you're using any part of the code for your research:\r\n\r\n \t- Narayanan, B. N., Hardie, R. C., De Silva, M. S., & Kueterman, N. K. (2020). Hybrid machine learning architecture for automated detection and grading of retinal images for diabetic retinopathy<\/a>. Journal of Medical Imaging, 7(3), 034501.<\/li>\r\n \t
- Narayanan, B. N., De Silva, M. S., Hardie, R. C., Kueterman, N. K., & Ali, R. (2019). Understanding Deep Neural Network Predictions for Medical Imaging Applications. arXiv preprint arXiv:1912.09621<\/a>.<\/li>\r\n<\/ul>\r\nThe Kaggle blindness detection challenge dataset (APTOS 2019 Dataset<\/a>) contains separate training and testing cases. In this blog, we solely utilize the training dataset to study and estimate the performance. These images were captured at the Aravind Eye Hospital, India. The training dataset contains 3662 images marked into different categories (Normal, Mild DR, Moderate DR, Severe DR, and Proliferative DR) by expert clinicians. Note that, in this blog, we solely focus on detecting DR, you could find more details about our grading architecture in our paper<\/a>.<\/em>\r\n
<\/h6>\r\nGrouping Data by Category<\/h2>\r\nWe extract the labels from excel sheet and segregate the images into 2-folders as 'no' or 'yes' as we're solely focused on detecting DR in this blog. The helper code for splitting the data into categories is at the end of this post. <\/em>\r\n<\/h6>\r\nLoad the Database<\/h2>\r\n<\/h6>\r\nLet's begin by loading the database using imageDatastore<\/a>. It's a computationally efficient function to load the images along with its labels for analysis.\r\n<\/h6>\r\n% Two-class Datapath<\/span>\r\ntwo_class_datapath = 'Train Dataset Two Classes';\r\n\r\n% Image Datastore<\/span>\r\nimds=imageDatastore(two_class_datapath, ...\r\n 'IncludeSubfolders',true, ...\r\n 'LabelSource','foldernames');\r\n\r\n% Determine the split up<\/span>\r\ntotal_split=countEachLabel(imds)\r\n<\/pre>\r\n![\"\"](\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2020\/08\/second_1.png\")
\r\nVisualize the Images<\/h2>\r\nLet's visualize the images and see how images differ for each class. It would also help us determine the type of classification technique that could be applied for distinguishing the two classes. Based on the images, we could identify preprocessing techniques that would assist our classification process. We could also determine the type of CNN architecture that could be utilized for the study based on the similarities within the class and differences across classes. In this article, we implement transfer learning using inception-v3 architecture. You can read<\/a> our paper to see the performance of different preprocessing operations and other established architectures.\r\n<\/h6>\r\n% Number of Images<\/span>\r\nnum_images=length(imds.Labels);\r\n\r\n% Visualize random 20 images<\/span>\r\nperm=randperm(num_images,20);\r\nfigure;\r\nfor idx=1:20\r\n \r\n subplot(4,5,idx);\r\n imshow(imread(imds.Files{perm(idx)}));\r\n title(sprintf('%s',imds.Labels(perm(idx))))\r\n \r\nend<\/pre>\r\n<\/h6>\r\n![\"\"](\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2020\/08\/retinopathy_2.png\")
\r\n<\/h6>\r\nTraining, Testing and Validation<\/h2>\r\nLet\u2019s split the dataset into training, validation and testing. At first, we are splitting the dataset into groups of 80% (training & validation) and 20% (testing). Make sure to split equal quantity of each class.\r\n<\/h6>\r\n% Split the Training and Testing Dataset<\/span>\r\ntrain_percent=0.80;\r\n[imdsTrain,imdsTest]=splitEachLabel(imds,train_percent,'randomize');\r\n \r\n% Split the Training and Validation<\/span>\r\nvalid_percent=0.1;\r\n[imdsValid,imdsTrain]=splitEachLabel(imdsTrain,valid_percent,'randomize');\r\n<\/pre>\r\nThis leaves us with the following counts:\r\n\r\n \r\n
<\/h6>\r\nDiabetic Retinopathy (DR) is one of the leading cause for blindness, affecting over 93 million people across the world. DR is an eye disease associated with diabetes. Detection and grading DR at an early stage would help in preventing permanent vision loss. Automated detection and grading during the retinal screening process would help in providing a valuable second opinion. In this blog, we implement a simple transfer-learning based approach using a deep Convolutional Neural Network (CNN) to detect DR<\/strong>.\r\n<\/h6>\r\nPlease cite the following articles if you're using any part of the code for your research:\r\n\r\n \t- Narayanan, B. N., Hardie, R. C., De Silva, M. S., & Kueterman, N. K. (2020). Hybrid machine learning architecture for automated detection and grading of retinal images for diabetic retinopathy<\/a>. Journal of Medical Imaging, 7(3), 034501.<\/li>\r\n \t
- Narayanan, B. N., De Silva, M. S., Hardie, R. C., Kueterman, N. K., & Ali, R. (2019). Understanding Deep Neural Network Predictions for Medical Imaging Applications. arXiv preprint arXiv:1912.09621<\/a>.<\/li>\r\n<\/ul>\r\nThe Kaggle blindness detection challenge dataset (APTOS 2019 Dataset<\/a>) contains separate training and testing cases. In this blog, we solely utilize the training dataset to study and estimate the performance. These images were captured at the Aravind Eye Hospital, India. The training dataset contains 3662 images marked into different categories (Normal, Mild DR, Moderate DR, Severe DR, and Proliferative DR) by expert clinicians. Note that, in this blog, we solely focus on detecting DR, you could find more details about our grading architecture in our paper<\/a>.<\/em>\r\n
<\/h6>\r\nGrouping Data by Category<\/h2>\r\nWe extract the labels from excel sheet and segregate the images into 2-folders as 'no' or 'yes' as we're solely focused on detecting DR in this blog. The helper code for splitting the data into categories is at the end of this post. <\/em>\r\n<\/h6>\r\nLoad the Database<\/h2>\r\n<\/h6>\r\nLet's begin by loading the database using imageDatastore<\/a>. It's a computationally efficient function to load the images along with its labels for analysis.\r\n<\/h6>\r\n% Two-class Datapath<\/span>\r\ntwo_class_datapath = 'Train Dataset Two Classes';\r\n\r\n% Image Datastore<\/span>\r\nimds=imageDatastore(two_class_datapath, ...\r\n 'IncludeSubfolders',true, ...\r\n 'LabelSource','foldernames');\r\n\r\n% Determine the split up<\/span>\r\ntotal_split=countEachLabel(imds)\r\n<\/pre>\r\n\r\nVisualize the Images<\/h2>\r\nLet's visualize the images and see how images differ for each class. It would also help us determine the type of classification technique that could be applied for distinguishing the two classes. Based on the images, we could identify preprocessing techniques that would assist our classification process. We could also determine the type of CNN architecture that could be utilized for the study based on the similarities within the class and differences across classes. In this article, we implement transfer learning using inception-v3 architecture. You can read<\/a> our paper to see the performance of different preprocessing operations and other established architectures.\r\n<\/h6>\r\n% Number of Images<\/span>\r\nnum_images=length(imds.Labels);\r\n\r\n% Visualize random 20 images<\/span>\r\nperm=randperm(num_images,20);\r\nfigure;\r\nfor idx=1:20\r\n \r\n subplot(4,5,idx);\r\n imshow(imread(imds.Files{perm(idx)}));\r\n title(sprintf('%s',imds.Labels(perm(idx))))\r\n \r\nend<\/pre>\r\n<\/h6>\r\n\r\n<\/h6>\r\nTraining, Testing and Validation<\/h2>\r\nLet\u2019s split the dataset into training, validation and testing. At first, we are splitting the dataset into groups of 80% (training & validation) and 20% (testing). Make sure to split equal quantity of each class.\r\n<\/h6>\r\n% Split the Training and Testing Dataset<\/span>\r\ntrain_percent=0.80;\r\n[imdsTrain,imdsTest]=splitEachLabel(imds,train_percent,'randomize');\r\n \r\n% Split the Training and Validation<\/span>\r\nvalid_percent=0.1;\r\n[imdsValid,imdsTrain]=splitEachLabel(imdsTrain,valid_percent,'randomize');\r\n<\/pre>\r\nThis leaves us with the following counts:\r\n\r\n \r\n
- \r\n \t
- Narayanan, B. N., Hardie, R. C., De Silva, M. S., & Kueterman, N. K. (2020). Hybrid machine learning architecture for automated detection and grading of retinal images for diabetic retinopathy<\/a>. Journal of Medical Imaging, 7(3), 034501.<\/li>\r\n \t
- Narayanan, B. N., De Silva, M. S., Hardie, R. C., Kueterman, N. K., & Ali, R. (2019). Understanding Deep Neural Network Predictions for Medical Imaging Applications. arXiv preprint arXiv:1912.09621<\/a>.<\/li>\r\n<\/ul>\r\nThe Kaggle blindness detection challenge dataset (APTOS 2019 Dataset<\/a>) contains separate training and testing cases. In this blog, we solely utilize the training dataset to study and estimate the performance. These images were captured at the Aravind Eye Hospital, India. The training dataset contains 3662 images marked into different categories (Normal, Mild DR, Moderate DR, Severe DR, and Proliferative DR) by expert clinicians. Note that, in this blog, we solely focus on detecting DR, you could find more details about our grading architecture in our paper<\/a>.<\/em>\r\n
<\/h6>\r\n
Grouping Data by Category<\/h2>\r\nWe extract the labels from excel sheet and segregate the images into 2-folders as 'no' or 'yes' as we're solely focused on detecting DR in this blog. The helper code for splitting the data into categories is at the end of this post. <\/em>\r\n
<\/h6>\r\n
Load the Database<\/h2>\r\n
<\/h6>\r\nLet's begin by loading the database using imageDatastore<\/a>. It's a computationally efficient function to load the images along with its labels for analysis.\r\n
<\/h6>\r\n
% Two-class Datapath<\/span>\r\ntwo_class_datapath = 'Train Dataset Two Classes';\r\n\r\n% Image Datastore<\/span>\r\nimds=imageDatastore(two_class_datapath, ...\r\n 'IncludeSubfolders',true, ...\r\n 'LabelSource','foldernames');\r\n\r\n% Determine the split up<\/span>\r\ntotal_split=countEachLabel(imds)\r\n<\/pre>\r\n\r\nVisualize the Images<\/h2>\r\nLet's visualize the images and see how images differ for each class. It would also help us determine the type of classification technique that could be applied for distinguishing the two classes. Based on the images, we could identify preprocessing techniques that would assist our classification process. We could also determine the type of CNN architecture that could be utilized for the study based on the similarities within the class and differences across classes. In this article, we implement transfer learning using inception-v3 architecture. You can read<\/a> our paper to see the performance of different preprocessing operations and other established architectures.\r\n
<\/h6>\r\n
% Number of Images<\/span>\r\nnum_images=length(imds.Labels);\r\n\r\n% Visualize random 20 images<\/span>\r\nperm=randperm(num_images,20);\r\nfigure;\r\nfor idx=1:20\r\n \r\n subplot(4,5,idx);\r\n imshow(imread(imds.Files{perm(idx)}));\r\n title(sprintf('%s',imds.Labels(perm(idx))))\r\n \r\nend<\/pre>\r\n<\/h6>\r\n
\r\n<\/h6>\r\n
Training, Testing and Validation<\/h2>\r\nLet\u2019s split the dataset into training, validation and testing. At first, we are splitting the dataset into groups of 80% (training & validation) and 20% (testing). Make sure to split equal quantity of each class.\r\n
<\/h6>\r\n
% Split the Training and Testing Dataset<\/span>\r\ntrain_percent=0.80;\r\n[imdsTrain,imdsTest]=splitEachLabel(imds,train_percent,'randomize');\r\n \r\n% Split the Training and Validation<\/span>\r\nvalid_percent=0.1;\r\n[imdsValid,imdsTrain]=splitEachLabel(imdsTrain,valid_percent,'randomize');\r\n<\/pre>\r\nThis leaves us with the following counts:\r\n\r\n \r\n - Narayanan, B. N., De Silva, M. S., Hardie, R. C., Kueterman, N. K., & Ali, R. (2019). Understanding Deep Neural Network Predictions for Medical Imaging Applications. arXiv preprint arXiv:1912.09621<\/a>.<\/li>\r\n<\/ul>\r\nThe Kaggle blindness detection challenge dataset (APTOS 2019 Dataset<\/a>) contains separate training and testing cases. In this blog, we solely utilize the training dataset to study and estimate the performance. These images were captured at the Aravind Eye Hospital, India. The training dataset contains 3662 images marked into different categories (Normal, Mild DR, Moderate DR, Severe DR, and Proliferative DR) by expert clinicians. Note that, in this blog, we solely focus on detecting DR, you could find more details about our grading architecture in our paper<\/a>.<\/em>\r\n
![\"\"](\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2020\/08\/retinopathy_3.png\")
\r\n![\"\"](\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2020\/08\/retinopathy_4a.png\")
\r\n![\"\"](\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2020\/08\/retinopathy_5.png\")
\r\n![\"\"](\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2020\/08\/retinopathy_6.png\")
\r\n![\"\"](\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2019\/10\/image_bio-1024x683.jpg\")
<\/td>\r\n