Title: Interest point detection using the superpixel segmentation and a binary descriptor
Abstract: Extracting interest points is one of the most important issues in techniques such as an object recognition and classification or a place recognition. Correct feature extraction can efficiently find out which parts of the image are likely to be unique, and can also match similar or identical parts in two different images. However, previous feature extraction algorithms typically extract features based on corner points with a large gradient, which results in degraded performance in less textured images, such as flat walls. In this paper, we propose a novel feature extraction method that can extract points in the texture-less images. Our method extracts interest points from the contour region of neighboring superpixels, which has a little or more changes of color values, then use a BRIEF descriptor to extract low-dimensional descriptors. Experimental results show that our method can extract features from texture-less regions and perform the matching between two different images with a high accuracy.
Publication Year: 2018
Publication Date: 2018-08-17
Language: en
Type: article
Indexed In: ['crossref']
Access and Citation
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot