Object Segmentation Using Machine Learning and Computer Vision Technique

EasyChair Preprint 10084, version 2

Abstract

In computer vision, the process of separating objects from their backgrounds in still or moving images is known as object segmentation. Due to the varying item looks, illumination, and occlusions, this work is difficult. Machine learning approaches have made considerable strides in object segmentation in recent years, delivering cutting-edge results on a number of benchmark datasets. This article presents an overview of current developments in computer vision and machine learning methods for object segmentation. In order to segment objects, standard computer vision techniques, including thresholding, regionbased segmentation, and edge detection, are first reviewed. We discuss a range of popular deep learning techniques, such as convolutional network, Mask R-CNN, and U-Net, exploring their benefits and drawbacks. lastly, we analyze the current progress in object segmentation, from semisupervised and unsupervised segmentation approaches to video object segmentation and domain adaptation. Additionally, we consider some of the unresolved challenges in object segmentation, including the issues of dealing with small objects, occlusions, and adapting to new domains.

Keyphrases: Artificial Intelligence (AI), Convolutional Neural Networks CNN, deep learning, machine learning, neural networks

@booklet{EasyChair:10084,
  author    = {Vaishali Aggarwal and Muzahid Islam and Muzammil Azeem and Chetan Mehta and Sangeeta Soni},
  title     = {Object Segmentation Using Machine Learning and Computer Vision Technique},
  howpublished = {EasyChair Preprint 10084},
  year      = {EasyChair, 2023}}