Skip to main content
Back to Projects
Computer Vision / OCRJun 2023 → Dec 2023

Cheque Processing & Digitization Ecosystem

End-to-end cheque digitization pipeline using OpenCV for image preprocessing, PaddleOCR for text and amount extraction, and a Flask + Docker orchestration layer.

PythonOpenCVPaddleOCRFlaskDockerPostgreSQL
Field accuracy
96.2%
Throughput
120 img/s
p95 latency
850ms

Problem & Solution

The Problem

Manual cheque processing is slow, error-prone, and expensive at scale. Banks need automated extraction of payee, amount, date, and signature region from cheque images with high accuracy.

The Solution

Built a Flask + Docker service that takes a cheque image, runs OpenCV preprocessing (deskew, denoise, contrast), then routes crops to PaddleOCR for payee, amount, date, and signature region extraction. Outputs are validated and stored as structured JSON.

System Architecture

End-to-end flow from intake to outcome

  1. 01

    Image Preprocessing

    OpenCV

    OpenCV deskews, denoises, and normalizes contrast on the cheque image.

  2. 02

    Field Localization

    OpenCV

    Heuristics localize payee, amount, date, and signature regions.

  3. 03

    OCR Extraction

    PaddleOCR

    PaddleOCR extracts text from each region.

  4. 04

    Validation

    Python

    Regex + amount-parsing rules validate the extracted fields.

  5. 05

    Orchestration

    FlaskDocker

    Flask + Docker expose the pipeline as a service.

Engineering Challenges & Wins

Optimization results and engineering tradeoffs

Optimization Results

  • OpenCV preprocessing (deskew, denoise, contrast) for clean OCR input.
  • PaddleOCR extraction of payee, amount, date, and signature region.
  • Flask + Docker for full pipeline orchestration.

Engineering Challenges

  • Handling handwritten and printed fields with a single OCR model.
  • Robust preprocessing for noisy, skewed real-world cheque scans.

More case studies

Explore other production systems I've engineered end-to-end.

All Projects

Related Projects

Computer Vision / OCR
Contextual AI & Medical Vision Pipelines

Suite of context-aware document intelligence pipelines: PDF intelligence (PyMuPDF + spaCy + BERT), passport MRZ segmentation, and prescription OCR.

PDF entity F1
0.91
MRZ accuracy
98.5%
Rx field accuracy
94%
PythonPyMuPDFspaCyBERTOpenCV+2
View Case Study
Autonomous AI AgentsFeatured
OmniShield ICES: Asynchronous AI Email Security Middleware

Mail-server gateway (Postfix Milter) that intercepts every inbound message, runs a fine-tuned BERT core (M-BSCE) through an asynchronous inference pipeline (RabbitMQ + Celery), and adds multimodal threat detection (PaddleOCR for Quishing) with SOAP clawbacks to the upstream mail system.

Quishing detection
98%+
p95 classification
<400ms
Async throughput
5k msg/min
PythonPostfix MilterRabbitMQCeleryBERT+6
View Case Study