Machine Learning
CXR Processing AI Engine
Chest X-ray Analysis & Report Generation Platform
A web app that allows users to upload chest X-ray images in DICOM, SVS, PNG, TIFF, and other formats, analyze findings, classify diseases, detect lung opacity, generate Grad-CAM outputs, and produce AI-generated medical reports.
Timeline
2024
Service
AI Product Engineering
Project Overview
- 1Uploads and processes chest X-ray images from multiple formats
- 2Classifies up to 14 disease categories from CXR scans
- 3Detects opacity regions and visualizes attention with Grad-CAM
- 4Generates report text based on imaging findings for fast review
Core Capabilities
AI / Pipeline Highlights
ViT Classification
Vision Transformer pipeline for disease category prediction.
YOLO Detection
Detection and localization workflow for opacity findings.
Report Generation
Transformer-based narrative generation for automated reporting.
Medical imaging report dashboard with Grad-CAM, findings, and generated narrative.
Tech Stack
Platform Value
Faster CXR analysis and triage
Shortens review cycles by combining classification, detection, and report output in one workflow.
Better interpretability
Grad-CAM and detection outputs make model behavior easier to review.
Scalable report workflow
Web-based processing supports repeatable report generation across batches.