Crate ddddocr 

ddddocr - Rust implementation of OCR for captcha recognition

This library is a Rust port of the Python ddddocr library, designed for offline local captcha recognition. It uses ONNX Runtime for inference and supports recognition of various captcha types including text and character-based captchas.

Features

• Offline local recognition - no network calls required
• Support for various captcha types
• Based on deep learning models trained on synthetic data
• Simple API with minimal dependencies
• Async inference support

Quick Start

use ddddocr::DdddOcr;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Initialize the OCR with the model file
    let mut ocr = DdddOcr::new("ddddocr.onnx")?;

    // Read the captcha image
    let image_bytes = std::fs::read("captcha.png")?;

    // Perform recognition
    let result = ocr.classification(&image_bytes).await?;

    println!("Recognized text: {}", result);
    Ok(())
}

Error Handling

The library uses a custom error type DdddOcrError for all operations:

use ddddocr::{DdddOcr, DdddOcrError};

async fn recognize_captcha(image_data: &[u8]) -> Result<String, DdddOcrError> {
    let mut ocr = DdddOcr::new("ddddocr.onnx")?;
    ocr.classification(image_data).await
}

Structs

DdddOcr

Main OCR struct for captcha recognition.

Enums

DdddOcrError

Error type for ddddocr operations.