Comparison of accuracy for novel optical character recognition technique over K-nearest neighbor algorithm to extract text from image and video
Sandhya Muppala, T Devi. · 2025
The main aim of this research is the comparison of Novel Optical Character Recognition (NOCR) Technique with K-Nearest Neighbor (K-NN) algorithm to extract text from image and video for improving accuracy. To extract keywords from a single document or a collection of documents, utilize the accuracy calculation. This Extraction of text is done by Natural-Language-Processing using the Novel OCR algorithm with a sample size of 22 and the K-Nearest Neighbor algorithm with a sample size of 44 that gives the result of an 81.95% G-power value. The accuracy produced by the Novel Optical Character Recognition algorithm is 81.95%, which is higher than K-Nearest Neighbor algorithm’s accuracy of 78.32%. It shows that there is no statistical significance difference between the Novel OCR algorithm and K-NN algorithm with p=0.254 (p>0.05). Compared to the accuracy of K-NN, which is 78.32%, the Novel Optical Character Recognition method has a high accuracy rating of 81.95%. Discrete Cosine Transform, Fast Fourier Transform, K-Nearest Neighbor, Natural Language Processing, Novel Optical Character Recognition, Simple Logistic, Technology Transfer.