IDC prediction in breast cancer histopathology images using deep residual learning with an accuracy of 99.37% in a subset of images containing a total of 7,500 microscopic images. The model link – Chandra-cc/GitHub
Pneumonia is an infectious disease that inflames the air sacs in the lungs. It may occur in one or both the lungs. The air sacs may fill with fluid or pus, causing cough followed by phlegm or pus, fever, chills, and difficulty in breathing. The causes of pneumonia can be a variety of organisms such as bacteria, viruses, and fungi. A person suffering from pneumonia apart from suffering from difficulty in respiration can also suffer from other complications such as bacteremia, lung abscess, pleural effusion and among countless others. This model presents an efficient and enhanced method for diagnosis of pneumonia using deep residual learning along with separable convolutional networks which achieved an astonishing accuracy of 98.22%.
The images were CXR (Chest-X-ray images) in grayscale format obtained from the Mendeley CXR data available at kaggle Dataset Link
The model link – Chandra-cc/Github
Object Classification using VGG16 Network on cifar10 datasets from keras.
The images were imported from keras.datasets containing about 50000 training images with 10000 testing images.
The model was created with VGG16 as it’s base model and a Residual Separable Convolutional layer following it.
The optimizer was Adam, 50 epochs with a batch size of 16 and early stopping with the patience of 10 and lr_reduce with a factor of 0.1 and patience of 2 were applied to the network.
The training set was split into validation set and training set itself with a validation split = 0.1
The link of the Model – Chandra-cc/GitHub
The project involves object classification from Resnet50 on the cifar dataset. link – object_classifiacation