📦 PyTorch based visualization package for generating layer-wise explanations for CNNs.

Going deeper into Deep CNNs through visualization methods: Saliency maps, optimize a random input image and deep dreaming with Keras

Code for the paper : "Weakly supervised segmentation with cross-modality equivariant constraints", available at https://arxiv.org/pdf/2104.02488.pdf

First position in Gran Canary Datathon 2021

Heat Map 🔥 Generation codes for using PyTorch and CAM Localization Algorithm.

Deep Learning Breast MRI Segmentation and Classification

We will build and train a Deep Convolutional Neural Network (CNN) with Residual Blocks to detect the type of scenery in an image. In addition, we will also use a technique known as Gradient-Weighted Class Activation Mapping (Grad-CAM) to visualize the regions of the inputs and help us explain how our CNN models think and make decision.

Repository of the course project of CMU 16-824 Visual Learning and Recognition

Distinguishing Natural and Computer-Generated Images using Multi-Colorspace fused EfficientNet

PyTorch MobileNetV2 Stanford Cars Dataset Classification (0.85 Accuracy)

Intracerebral Hemorrhage Detection on Computed Tomography Images Using a Residual Neural Network

Develop and train image classification models using advanced deep learning techniques to identify diseases specific to apples.

Using LIME and Grad-CAM techniques to explain the results achieved by various image transfer learning techniques

This study tries to compare the detection of lung diseases using xray scans from three different datasets using three different neural network architectures using Pytorch and perform an ablation study by changing learning rates. The dimensional understanding is visualised using t-SNE and Grad-CAM for visualisation of diseases in x-ray scans.

Collecting fish image data, after training classifiers grad-cam is applied for the prediction interpretation

Exploring the Application of Attention Mechanisms in Conjunction with Baseline Models on the COVID-19-CT Dataset

Gradient Frequency Attention: Tell Neural Networks where speaker information is.

KL severity grading using SE-ResNet and SE-DenseNet architectures trained with Cross Entropy loss and Focal Loss. The hyperparameters of focal loss have been fine-tuned as well. Further, Grad-CAM has been implemented for visualization purposes.

Have you ever asked yourself, which regions of the input image were considered more by the model? If so, Grad-CAM has exciting answers for you!

Distinguishing Natural and Computer-Generated Images using Multi-Colorspace fused EfficientNet