Fast SSOCR: A12E Weight Scale Reader

A high-performance, lightweight FastAPI application designed to extract and format numerical readings from A12E LED weight scale displays.

📸 Demo

This project is specifically optimized to handle the 7-segment red LED output of the A12E indicator, accurately identifying weights even with glare or background noise.

Case A (179.5 kg)	Case B (99.5 kg)	Case C (94.0 kg)

🚀 Overview

This project provides a specialized OCR pipeline for 7-segment LED displays. It uses advanced image processing (HSV Red-Dominance Masking) to isolate glowing LED digits from complex backgrounds, followed by EasyOCR for high-accuracy character recognition.

✨ Features

• Red-Dominance Masking: Isolates glowing red LED segments while ignoring shadows, reflections, and scale casing.
• Morphological Refinement: Dilation and closing operations to bridge gaps in 7-segment characters.
• Precision Formatting: Automatically formats raw OCR output to match A12E standards (e.g., 1795 -> 179.5 kg).
• FastAPI Backend: Ready-to-deploy RESTful API with automated Swagger documentation.
• CPU Optimized: Designed to run efficiently on standard hardware without requiring a GPU.

🛠️ Technology Stack

• FastAPI: Modern, high-performance web framework.
• OpenCV: Image processing (Color masking, Dilation, Pre-processing).
• EasyOCR: Robust optical character recognition.
• NumPy: Array processing for high-speed image manipulation.

📋 Installation

1. Clone the repository:

   git clone [https://github.com/limkhysok/fast-ssocr.git](https://github.com/limkhysok/fast-ssocr.git)
   cd fast-ssocr

2. Install dependencies:

   pip install -r requirements.txt

3. Run the application:

   python main.py

🔌 API Usage

Extract Weight

Endpoint: POST /extract-weight/

Request:

• file: Image file (JPG/PNG) of the weight scale display.

Response:

{
  "raw_ocr": "1795",
  "formatted_weight": "179.5",
  "unit": "kg"
}