International Journal of Advances in Electrical Engineering
2024, Vol. 5, Issue 1, Part B
Handwritten formula recognition using CNN
Author(s): Rashmi Deshpande
Abstract: This paper presents a machine intelligence approach for recognizing handwritten formulas. The process involves three steps: acquiring data, training data, extracting features using a Convolution Neural Network, and matching the features and calculating the probability. The Convolution Neural Network is used to improve the data rate of recognition of handwritten formula symbols, compensating for information loss in the formula. The experimental results show that the Convolution Neural Network is effective for feature extraction and improving the data recognition rate. The paper also proposes an alternative method for converting handwritten mathematical formulas into computer-readable text using horizontal and vertical projections and convolution neural networks. This method successfully segmented and identified each character in handwritten equations from lined papers, allowing them to be converted into a digital format for data processing like editing. This method is time-efficient.
Pages: 95-98 | Views: 74 | Downloads: 56
Download Full Article: Click Here
How to cite this article:
Rashmi Deshpande. Handwritten formula recognition using CNN. Int J Adv Electr Eng 2024;5(1):95-98.
