ICDAR 2026 Competition on Writer Identification and Pen Classification from Hand-Drawn Circles

2026-05-08

AI Summary

This study addresses the novel challenge of jointly identifying writer identity and pen type from minimal static handwriting—specifically, a single hand-drawn circle—thereby exploring the coupling mechanism between biometric traits and physical tool traces. To this end, the authors construct a large-scale dataset comprising 46,155 scanned images at 400 DPI and define two core tasks: open-set writer identification and cross-writer pen classification. A ResNet-based deep learning approach achieves 92.726% Top-1 accuracy in pen classification and 64.801% in writer identification. Through a Kaggle competition attracting over 500 participants and nearly 5,000 submissions, the work establishes a new benchmark for minimal handwriting trace analysis and demonstrates the feasibility of disentangling and extracting dual information from a single circular stroke.

Abstract

This paper presents CircleID, a large-scale ICDAR 2026 competition on writer identification and pen classification from scanned hand-drawn circles. The primary objective is to investigate how biometric writer characteristics and physical pen features naturally entangle within minimal, static traces. CircleID comprises two distinct tasks: (1) open-set writer identification, requiring models to recognize known writers while explicitly rejecting unknown ones, and (2) cross-writer pen classification, evaluated across both seen and unseen writers. Participants were provided with a new, controlled dataset of 46,155 tightly cropped circle images, digitized at 400 DPI and annotated for writer identity and pen type. The dataset comprises samples from 50 known and 16 unknown writers using eight different pens. Hosted on Kaggle as two separate tracks with public and private leaderboards, the competition provided participants with a ResNet baseline. In total, 389 teams (436 participants) made 3,185 submissions for the pen classification task, and 113 teams (141 participants) made 1,737 submissions for the writer identification track. The best-performing private leaderboard submissions achieved a Top-1 accuracy of 64.801% for writer identification and 92.726% for pen classification. This paper details the dataset, evaluates the winning methodologies, and analyzes the impact of out-of-distribution writers on model generalization and feature disentanglement. In this large-scale competition, CircleID establishes a new baseline for minimal-trace analysis.

Problem

• writer identification
• pen classification
• hand-drawn circles
• biometric characteristics
• minimal-trace analysis

Innovation

• writer identification
• pen classification
• minimal-trace analysis
• feature disentanglement
• open-set recognition