A Comparative Study of Local and Global Feature Detector and Descriptors for Image and Video Retrieval

Abstract

AbstractIn this paper, we are studying about two leading approaches. Step one, we will define the local features detection and description method for classifying pattern recognition, image tracking, clustering object recognition, or identification images based on local image patches or key points in the images of an object. Secondly, we can define global feature detection and description methods for an image that are used in image retrieval and object detection for a whole image. Local and global feature methods are comprehensive, and their output is compared using a dataset. The images were selected according to the hypothesis that they could be better described using global features. Selecting the algorithms, that all depends on the feature’s detection and description for the data. This led to a detailed review of local and global video or image detectors and new descriptors. We presented a summary of existing performance evaluation and databases of benchmarks. Lastly, in terms of future directions, we have completed the investigation. This research study may assist and provide a framework for image recognition/processing and machine learning.KeywordsLocal feature detector and descriptorGlobal feature detector and descriptorMatching patternBenchmark datasetsVideo copy detection