Motion-Based Handwriting Recognition and Word Reconstruction
This addresses handwriting recognition for applications like digital pens, but it appears incremental as it builds on existing methods without major breakthroughs.
The paper tackles the problem of recognizing handwritten words from continuous motion data by using a single-letter classifier and a reconstruction pipeline with dynamic programming and auto-correction. It achieves results through model optimization and domain adaptation for unseen data distributions, but no concrete numbers are provided.
In this project, we leverage a trained single-letter classifier to predict the written word from a continuously written word sequence, by designing a word reconstruction pipeline consisting of a dynamic-programming algorithm and an auto-correction model. We conduct experiments to optimize models in this pipeline, then employ domain adaptation to explore using this pipeline on unseen data distributions.