project-proposal-2025

ScanNote – Quick Summaries of Handwritten Notes

Abstract

ScanNote is a lightweight web/mobile application that digitizes handwritten notes and auto-generates concise summaries of their content. By integrating optical character recognition (OCR) with an AI-based text summarization service, the system allows users to quickly capture images of their notes, store them in the cloud, and retrieve a short, structured summary. ScanNote aims to reduce the hassle of managing scattered paper notes and speed up the review process for students and professionals.

Author

Name: Stanley Chen
Student Number: 48287803

Functionality

1. Note Capture & Storage
   • Users photograph or upload scanned images of handwritten notes through a mobile or web interface.
   • The images are stored securely in the cloud with minimal metadata (title, timestamp, optional tags).
2. OCR & Summarization
   • The system applies OCR to extract text from the note images.
   • An AI summarization component condenses the recognized text into bullet points, a short title, and a brief paragraph capturing main ideas.
3. Browse & Retrieve
   • Each note entry contains both the original image and a condensed summary.
   • Users can quickly view or edit the AI-generated summary if needed; basic search by tags or keywords is provided in the MVP.

Scope

• MVP
  • Single-page or single-image note uploads in one language (e.g., English).
  • Basic user authentication and note organization by date or tag.
  • Summaries focus on straightforward text (no complex diagrams or multi-page documents).
• Out of Scope
  • Advanced handwriting style detection, multi-language OCR, or real-time collaboration.
  • Table or formula parsing.

Quality Attributes

1. Accuracy
   • Target: Achieve ~80% OCR accuracy for clear handwriting, and user satisfaction of 70%+ for summary relevance.
   • Approach: Utilize widely used OCR libraries and an off-the-shelf summarization model (e.g., transformer-based API).
2. Reliability
   • Goal: Ensure note processing and retrieval remain stable under typical usage.
   • Target: 99% service uptime, average note processing time under 10 seconds.
   • Implementation: Cloud hosting with basic logging, error handling, and retry logic.
3. Security & Privacy
   • Focus: Users’ handwritten notes can be sensitive.
   • Objective: All uploads and summaries accessible only by the owner, using encrypted storage and HTTPS for data transmission.

Evaluation

1. Accuracy Checks
   • Collect a small set of handwritten samples; measure OCR accuracy by comparing recognized text with ground truth.
   • Evaluate summarization clarity via user feedback questionnaires.
2. Reliability Monitoring
   • Measure uptime via a simple server monitoring tool over a week/month.
   • Check average processing time with moderate concurrent usage (~20 users).
3. Security Testing
   • Verify that only authenticated users can access their own notes.
   • Inspect data encryption settings in the hosting environment, ensuring compliance with basic privacy standards.