Performance comparison of segmentation algorithms for hand gesture recognition

Abstract

The gestures presented in diverse backgrounds have to be accurately processed and segmented, for it to be classified precisely by the hand gesture recognition system. This study compares performance of the proposed Image Segmentation Algorithm with a standard Canny Edge Detection Algorithm by comparing the statistical values of the features obtained from the feature extraction stage, thus validating the importance of having a robust preprocessing stage for the hand gestures. The proposed algorithm uses Non-local Mean filter for noise removal and then an improved Global Swarm Optimization based Canny edge detection for extracting the edges. Features are extracted using two dimensional Multi-resolution Discrete Wavelet Transform (2D-DWT) combined with Gray-level Co-occurrence Matrix. The efficiency of the proposed Image Segmentation Algorithm is evaluated using Radial Basis Function Neural Network as the classifier.

Forecasting of Breast Cancer Using Voting Classifier

“…The dataset used for analysis may contain inconsistencies like missing values, outliers and it has to be handled before being used to build the model. After the implementation of all the algorithms with the information provided the result is determined based on the accuracy of the used algorithms [1][2].…”

“…2 shows confusion matrix which describes the visualization of the performance of a classification model on a test data.It provides a summary of number of correct and incorrect predictions with count values and provide best accuracy among these algorithms .…”