How dhilipsanjay/detection-of-sensitive-data-exposure-in-images works
This repository provides a web application for warning users when images contain sensitive information. The documented workflow combines image classification with OCR-derived text classification, then fuses both results into a sensitivity decision. The README describes these stages but no source excerpts were available, so the graph uses documented relationships only; implementation details such as where fusion occurs are uncertain.
Main entry points: script.js Classification workflow · index.html Web interface · image_model/model.json Image classifier · text_model/model.json Text classifier
Components 9
Web application
- Web interface
webapp/public/index.html - Classification workflow
webapp/public/assets/script.js - Sensitivity result
webapp/public/index.html
Image and text analysis
- Image classifier
webapp/public/models/image_model/model.json - Tesseract OCR
- Text classifier
webapp/public/models/text_model/model.json - Result fusion
webapp/public/assets/script.js
Model artifacts
- Text word index
webapp/public/models/text_model/word_index.json
Other components
- User
Connections 9
- User → Web interface: submits image
README.md - Web interface → Classification workflow: starts analysis
README.md - Classification workflow → Image classifier: classifies image
README.md - Classification workflow → Tesseract OCR: extracts text
README.md - Classification workflow → Text classifier: classifies text
README.md - Classification workflow → Text word index: uses vocabulary
- Classification workflow → Result fusion: combines results
README.md - Result fusion → Sensitivity result: produces decision
README.md - Sensitivity result → User: warns user
README.md