Open AccessTracking and Detection of Vehicles using Locality Sensitive Histogram (LSH) Feature Extraction
Author(s) -
Bhavya Rudraiah*,
Geetha K. S.
Publication year - 2020
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.e3048.039520
Subject(s) - artificial intelligence , histogram , computer science , computer vision , pattern recognition (psychology) , locality , false positive paradox , support vector machine , classifier (uml) , feature extraction , bin , object detection , image (mathematics) , algorithm , philosophy , linguistics
Detection and tracking has become a vital chore in most of the computer vision applications. It analyzes the behavior of the object and detects when it appears in other frames. In this paper, a locality sensitive histogram (LSH) algorithm along with SVM is used to detect and track the objects. Locality Sensitive Histogram is used for feature extraction and detection. It is computed at each pixel location, by adding a floating-point value to bin, which is its unique nature. The extracted features are subjected to Linear SVM classifier and then the object is tracked by eliminating false positives. This proposed method precisely tracks and detects the object well with different challenges. Experimental results demonstrate the performance of the proposed algorithm with an accuracy of 89% considering several challenging factors. Evaluation of various other algorithms using different performance parameters is also tabulated in the diagram and shows that the proposed method is topmost performer in tracking the objects. This method can be utilized to track different objects of different scale and track efficiently.