A DATASET OF KAZAKH HANDWRITTEN CHARACTERS

Abstract

Handwriting still appears in classrooms, paperwork, and personal records, even where most communication has moved to a screen. For Kazakh written in Cyrillic, the main blocker for character-level recognition research is not a lack of ideas but a lack of carefully described data that other groups can reuse with confidence. This paper treats data quality as the central engineering problem and walks through how a usable dataset was put together. We describe form design and writer recruitment, scanning at 300 dpi, semi-automatic segmentation of grid forms into character crops, normalization and resizing to 64×64, and a two-pass annotation workflow with class-wise spot checks. The resulting dataset holds approximately 140,000 grayscale images across 42 classes from around 110 writers in two age groups. We also report a baseline convolutional neural network that reaches 84.3 % test accuracy and a macro-averaged F1 of 0.836, mainly to confirm the dataset is usable and to give later work a comparable reference point. The contribution is the methodology more than the size: per-class auditing, structured handling of segmentation failures, and a strict ambiguity-removal rule during labeling did more for reliability than additional samples would have. Keywords: handwritten character dataset, Kazakh Cyrillic, data quality, segmentation, annotation workflow, label noise, convolutional neural network, low-resource scripts.

More by Dinara Kaibassova

• A Reproducible Hybrid AI Framework for Early Soil Nutrient Screening from Sentinel-2 Remote Sensing DataTechnologies · 2026
• Evaluation Protocol for Morphology Tolerant Predictive Text in Kazakh Using N Gram ModelsIEEE International Conference on Smart Information Systems and Technologies · 2026
• INTERPRETING CLINICAL RESEARCH FOR DISEASE DIAGNOSIS: A MACHINE LEARNING PERSPECTIVEBulletin of Toraighyrov University Physics & Mathematics series · 2026
• COMPARATIVE ANALYSIS OF DEEP LEARNING MODELS FOR CHEST DISEASE DIAGNOSIS USING NIH X-RAY DATASETScientific Journal of Astana IT University · 2026