PyTorch re-implementation of Grad-CAM (+ vanilla/guided backpropagation, deconvnet, and occlusion sensitivity maps)
-
Updated
Jan 21, 2020 - Python
PyTorch re-implementation of Grad-CAM (+ vanilla/guided backpropagation, deconvnet, and occlusion sensitivity maps)
Pytorch implementation of various neural network interpretability methods
This script evaluates the sensitivity of VGG-16 to occlusion using Keras
Visualizing and interpreting features of CNN model
IO/NIO-Bildklassifikation für die industrielle Qualitätsprüfung mit PyTorch, Transfer Learning und Explainable AI (Integrated Gradients). Entwickelt im Rahmen einer Bachelorarbeit.
A CNN model developed to observe and predict lung related chronic diseases at an early stage. ResNet-50 architecture is used to classify x-ray images and Occlusion Sensitivity Function is used to visualise the image with its affected areas.
Explainable AI in Intrusion Detection
Educational MICCAI tutorial on stress-testing saliency maps, shortcut learning, and trustworthy medical imaging AI.
Deep learning pipeline for automatic classification of ERCP fluoroscopic images using CNN models, threshold tuning, ensemble learning, TTA and interpretability methods such as Grad-CAM and Occlusion Sensitivity.
PyTorch comparison of MLP and CNN architectures for five-class blood-cell image classification, with feature maps, softmax error analysis, spatial robustness, and occlusion sensitivity.
To associate your repository with the occlusion-sensitivity topic, visit your repo's landing page and select "manage topics."