IJMTES-An Unsupervised Object Tracking And Detection Using Otsu Algorithm

Journal Title : International Journal of Modern Trends in Engineering and Science

Paper Title : An Unsupervised Object Tracking And Detection Using Otsu Algorithm

Author’s Name : Sundaresan K, Dinesh Kumar T

Volume 07 Issue 02  2020unnamed

ISSN no:  2348-3121

Page no: 01-05

Abstract –  Object detection is one of the major goals in computer vision that deals with detecting instances of semantic objects of a certain class. In Video acquisition takes video as an input and splits the number of frames from the given video. Then it removes the noise from the frame using pre-processing and segment the moving object using Otsu algorithm. Region Of Interest can be calculated to perform further analysis for image identification and recognition. Histogram of Oriented Gradients is a feature descriptor used for object detection and using Histogram of Oriented Gradients all the humans in the video can be detected. Random forest is a learning method for classification and regression and by matching the characteristic of the image using random forest classifier particular human can be detected in video. Finally, track the moving object in original video.Object can be detected and tracked based on Otsu algorithm and Random forest. Finally, it detects the image in the video and track the moving object in original video. Based on the characteristic matching of the two images tracking can be performed. Object tracking and detection can be implemented using Matrix Laboratory.

Keywords –  Image Processing, MATLAB, OTSU Segmentation, HOG Extraction.