HSMM-Based Offline Handwritten Mathematical Expression Recognition with Duration Modeling and Error Guarantees

Author(s)	Ms. JAYAPRIYA V, Prof. Dr. NITHYAPRIYA N
Country	India
Abstract	This paper proposes HSMM architecture for doing offline handwritten mathematical recognition. In HSMM, since the state duration models are considered, therefore the model performs better in segmentation and recognition of elongated and multi-stroked symbols. Through the use of HOG and through the optimization of algorithm using Baum Welch EM algorithm, HSMM outperforms HMM, CNN HMM, and Transformer CTC algorithms based on their Character Error Rate (CER), Word Error Rate (WER), and equation level accuracy. The comparative study demonstrates the error reduction capability of HSMM and the equation level study establishes that the model is robust in the case of calculus, algebra, trigonometry, and limit equations. The decrease in substitution, insertion and deletion errors is also verified using the error breakdown analysis. With the help of convergence, error bound, and computational efficiency property, HSMM has strong theoretical foundations for handwritten mathematical OCR problem.
Keywords	Hidden Semi Markov Model (HSMM), Handwritten Mathematical Recognition, Equation Recognition.
Field	Mathematics > Statistics
Published In	Volume 8, Issue 4, July-August 2026
Published On	2026-07-14
DOI	https://doi.org/10.36948/ijfmr.2026.v08i04.83716

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research