OCR Based Image Text to Speech Conversion using K-Nearest Neighbors and Comparing with Fuzzy K-Means Clustering Algorithm

Using a variety of various techniques, this research is to transform OCR-based visual text to spoken language. The image dataset was used for novel K-Nearest Neighbours (KNN) and fuzzy k-means clustering (FKM). The novel k-nearest neighbours (KNN) (N = 10) and the fuzzy k-means clustering (FKM) (N = 10) groups are compared to one another in this study. according to the results of a calculation made with the g power software, the total sample size was determined to be 80% powerful, with a confidence interval of 95%, an enrolment ratio of 0.1, and an alpha value of 0.05. The SPSS software as well as an independent sample t-test were used to conduct the comparison of the accuracy and specificity rates. The results of a MATLAB simulation indicate that fuzzy k-means clustering, also known as FKM, obtains an accuracy of 93 percent when applied to text to vocal translation. The statistical analysis done with SPSS managed to reach a significant level of accuracy (0.02). The newly proposed revolutionary k-nearest neighbours method performs better than the FKM method for voice converting using the provided pictures.