<- Back to projects

Vision-to-sentiment pipeline

SentiOCR

Full-stack AI platform that extracts text from handwritten images with OpenCV-preprocessed OCR and classifies emotional tone with fuzzy scoring.

PythonFastAPIOCROpenCVReact
SO
Public

Technical proof

OpenCV preprocessing
OCR stage
Sentiment classification
Open sourceDiscuss this style

Problem

Handwritten notes are difficult to search and analyze, especially when the emotional tone matters.

Solution

An OCR and sentiment pipeline that preprocesses images, extracts handwritten text, and scores emotional tone with a deterministic fuzzy system.

Architecture

System flow

Step 1

Image upload

Step 2

OpenCV preprocessing

Step 3

OCR extraction

Step 4

Fuzzy sentiment scoring

Step 5

React result UI

Stack

Python
FastAPI
OpenCV
OCR
React

Highlights

Vision plus language pipeline
Deterministic scoring for explainability
Full-stack result experience

Lessons

Preprocessing quality controls OCR quality
Transparent scoring can be more useful than opaque model output