Examining Dynamic Time Warping Algorithms for Hyper Spectral Image Data

Dynamic Time Warping (DTW) algorithms have currently been implemented to provide efficient and accurate time series analysis in a selection of programs. This paper examines the performance of numerous DTW algorithms within the context of hyperspectral image records. Specifically, the focal point of the research is to determine the most suitable technique for mapping and producing effective initialized fashions for use in imaging segmentation and feature extraction responsibilities. The software of those algorithms is verified by utilizing exclusive styles of hyperspectral imagery and appearing example-precise experiments for every one of the algorithms. The GLCM texture function set is used to have a look at modifications in the statistics set with respect to distinct alterations. The overall performance evaluation metrics used to evaluate the diverse techniques are time, accuracy, and inter-pixel as opposed to intra-pixel dependencies. Results show that the SOGA-DTW is the most appropriate alternative for interpreting the hyperspectral information sets in phrases of time, accuracy, and its ability to version inter-pixel and intra-pixel relationships. The changed DTW algorithms additionally look like greater stability and robustness over varying ameliorations. Finally, the ability to create a customized version for a specific instance of statistics is highlighted as a precious function while thinking about DTW for hyperspectral imagery analysis.