Segmentation of handwritten character identification from an image using digital image processing based on threshold level compared with edge based segmentation

The primary purpose of this research is to apply segmentation of handwritten identification from an image using digital image processing by utilising innovative edge detection segmentation in contrast with the Threshold level approach. The dataset that is used in this work makes use of the database that is maintained by the Computer Vision Lab at National TsingHua University (NTHU), which is accessible to the general public. The calculation is carried out using G-power 0.8, and the alpha and beta qualities are 0.05 and 0.2 with a confidence interval set at 95%. The sample size for the handwritten identification segmentation from an image using digital image processing with an improved accuracy rate was sample 280 (Group 1 =140 and Group 2 =140). Edge detection segmentation with a number of samples (N=10) and threshold level method with the same number of samples (N=10) are used to perform the segmentation of handwritten identification from an image using digital image processing with an improved accuracy rate. Both of these methods are used in the segmentation of handwritten identification from an image. When contrasted with the accuracy rate of the Threshold level approach, which is 90.823 percent, the accuracy rate of the Edge detection segmentation classifier is 93.82 seven times higher. The significance level of this investigation is p=0.0394, according to the findings. When compared to the Threshold level method for segmentation of handwritten character identification from an image using digital image processing, Edge detection segmentation yields significantly better results in terms of accuracy rate. This is because Edge detection segmentation uses more information from the original image.