Inspiration
Our inspiration for Maskerade came from the ever increasing need to protect privacy in a world where photos and documents are constantly shared online. We wanted to create a tool that automatically identifies and “masks” sensitive information such as faces, personal identifiers and confidential information without requiring any manual effort.
What it does
Maskerade automatically detects sensitive information in images, including faces, text (names, emails, credit card numbers), and other personally identifiable information (PII), and redacts them using AI. Users can upload an image and receive a “masked” version that protects their own and others’ privacy while keeping the image usable.
How we built it
- EasyOCR for extracting text from images
- face_detection library
- presidio for PII detection
Challenges we ran into
- Ensuring accurate detection of PII in text in images using OCR output
- Properly masking only the sensitive parts without affecting the rest of the image. We decided on using solid black rectangles instead of traditional blurring to guarantee that the sensitive information was fully obscured.
Accomplishments that we're proud of
- Built a fully functional prototype capable of automatically redacting images
What we learnt
- How to design privacy-first applications that give users control over sensitive data.
- The importance of UX/UI design for privacy tools like showing the user what is being redacted
What's next for Maskerade
We intend to add more features such as:
- Location privacy: Detecting geo-revealing objects like license plates, road signs, or landmarks and masking them
- Video and PDF support
- Customisable redaction modes
Log in or sign up for Devpost to join the conversation.