2xText2HD v1

Makes low-quality scans of text sharp and readable at 2×.

A RealPLKSR model for upscaling low-quality text, such as poor scans and photos of documents, into clear, legible pages. It works best on moderately sized text rather than very small print.

Download ONNX View repository Paper

Text enhancement Computer Vision

Scale: 2×
Parameters: 7.4M
CPU time per 256 px tile: 3.9s
ONNX file: 30MB

Example output

Drag the handle to compare. Produced by running this model.

After: Input: a 260 × 130 px low-quality scan of printed text. Output: the model's 2× result. — Input: a 260 × 130 px low-quality scan of printed text. Output: the model's 2× result.

Before: Input: a 260 × 130 px low-quality scan of printed text. Output: the model's 2× result. — Input: a 260 × 130 px low-quality scan of printed text. Output: the model's 2× result.

Technical details

Architecture: RealPLKSR (partial large kernel CNN)
Training data: Scanned books and text
Training: 168,000 iterations
ONNX build: FP32, opset 17, any image size
Input: RGB, NCHW, values 0–1
Benchmark: ONNX Runtime CPU, Apple M2 Pro
Original release: 27 August 2024

Good for

Scanned documents
Photos of receipts and pages
Screenshots of text

Run it in Flutter

This model runs as-is with our open-source flutter_super_resolution package (MIT), which tiles large images and uses Core ML on Apple devices when available. It works on iOS, Android (ARM devices), macOS and Windows.

pubspec.yaml · Darttext2xhigh.onnx

dependencies:
  flutter_super_resolution: ^1.1.1

// Load the model, then upscale a ui.Image
final upscaler = FlutterUpscaler(tileSize: 128, overlap: 8);
await upscaler.initializeModel('assets/models/text2xhigh.onnx');

final result = await upscaler.upscaleImage(
  image,
  scale: 2,
  onProgress: (progress, message) => print('$message ${(progress * 100).round()}%'),
);

upscaler.dispose();

Package on GitHub All models that work with it