wifi-densepose/examples/scipix/web

Scipix WASM - WebAssembly OCR

High-performance OCR with LaTeX support for the browser, powered by WebAssembly.

Features

• 📸 Image OCR: Recognize text from images
• 🧮 LaTeX Support: Extract mathematical formulas
• ⚡ Web Workers: Off-main-thread processing
• 🎯 TypeScript: Full type definitions
• 🚀 Optimized: <2MB bundle size
• 🔧 Flexible API: Multiple input formats

Quick Start

Installation

npm install ruvector-scipix-wasm

Build from Source

cd examples/scipix
npm run build

Basic Usage

import { createScipix } from 'ruvector-scipix-wasm';

// Initialize
const scipix = await createScipix();

// Recognize from file
const result = await scipix.recognize(imageData);
console.log(result.text);
console.log(result.latex);

Canvas Example

import { recognizeCanvas } from 'ruvector-scipix-wasm';

const canvas = document.getElementById('myCanvas');
const result = await recognizeCanvas(canvas);

Web Worker Example

import { createWorker } from 'ruvector-scipix-wasm';

const worker = createWorker();

// Process in background
const result = await worker.recognize(imageData);

// Batch processing with progress
const results = await worker.recognizeBatch(images, {
  onProgress: ({ processed, total }) => {
    console.log(`Progress: ${processed}/${total}`);
  }
});

worker.terminate();

API Reference

createScipix(options?)

Create a new Scipix instance.

const scipix = await createScipix({
  format: 'both',           // 'text' | 'latex' | 'both'
  confidenceThreshold: 0.5  // 0.0 - 1.0
});

ScipixWasm

Main API class.

Methods

• recognize(imageData: Uint8Array): Promise<OcrResult>
• recognizeFromCanvas(canvas: HTMLCanvasElement): Promise<OcrResult>
• recognizeBase64(base64: string): Promise<OcrResult>
• recognizeImageData(imageData: ImageData): Promise<OcrResult>
• recognizeBatch(images: Uint8Array[]): Promise<OcrResult[]>
• setFormat(format: RecognitionFormat): void
• setConfidenceThreshold(threshold: number): void
• getVersion(): string

Helper Functions

import {
  recognizeFile,      // From File/Blob
  recognizeCanvas,    // From HTMLCanvasElement
  recognizeBase64,    // From base64 string
  recognizeUrl,       // From image URL
  recognizeBatch,     // Batch processing
  imageToCanvas,      // Convert image to canvas
} from 'ruvector-scipix-wasm';

Types

OcrResult

interface OcrResult {
  text: string;           // Recognized text
  latex?: string;         // LaTeX (if enabled)
  confidence: number;     // 0.0 - 1.0
  metadata?: {
    width?: number;
    height?: number;
    format?: string;
  };
}

RecognitionFormat

type RecognitionFormat = 'text' | 'latex' | 'both';

Demo

Run the interactive demo:

npm run dev

Open http://localhost:8080/example.html

Performance Tips

1. Use Web Workers for large images or batch processing
2. Set confidence threshold to filter low-quality results
3. Resize images before processing if possible
4. Reuse instances instead of creating new ones
5. Use SharedImageBuffer for large image batches

Browser Support

• Chrome 57+
• Firefox 52+
• Safari 11+
• Edge 16+

Requires WebAssembly support.

Bundle Size

• WASM module: ~800KB (gzipped)
• JavaScript wrapper: ~15KB (gzipped)
• Total: <1MB

License

MIT

Credits

Built with:

• wasm-bindgen
• image-rs
• ruvector