Radiological Vision Suite
Deep learning transfer analysis of chest radiographs with target layer Grad-CAM visual attention mapping.
Image Validation Active
Radiograph Input
ResNet18 Backbone
Drag & drop chest X-ray image
DICOM, JPEG, PNG supported • Max 15MB
Only medical chest radiographs accepted
Model Architecture: ResNet-18 pretrained on ImageNet, fine-tuned on 5,856 chest radiographs with weighted CrossEntropyLoss for pathology localization.
Awaiting Radiograph Upload
Upload a chest radiograph or load a test sample on the left panel to execute deep residual analysis and Grad-CAM attention rendering.