nauman-riaz/DHSD: DHSD v1.0.0 — Initial Public Release

DHSD v1.0.0 — Initial Public Release This release introduces the German Handwriting Street-name Dataset (DHSD), a dataset designed for research in handwritten text recognition (HTR) for the German language. 📦 Contents 5,939 handwritten word images (PNG, 256×64, RGBA) Contributions from 37 writers Word instances derived from German city and street names Coverage of the complete German alphabet, including ä, ö, ü, ß CSV annotations linking each image to: Ground-truth transcription Writer ID Predefined dataset split: 80% Training 20% Testing 🎯 Purpose DHSD is intended to support: Word-level handwritten text recognition Writer-independent evaluation Research on German-language OCR / HTR systems Benchmarking deep learning models on structured lexical vocabularies 📚 Data Source Geographic names were derived from OpenStreetMap © OpenStreetMap contributors (ODbL). 📄 License Released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license to encourage open research use.