Abstract
Image classification, object detection and image semantic segmentation are three core research topics in computer vision. Among them, image semantic segmentation is the most challenging task. In actual segmentation, both the accuracy of segmentation and the speed of segmentation should be considered. The key to improve the accuracy is to determine the regions where different categories are located, that is to accurately define the edge contours of different categories. For this reason, considering on the difficulty of forecasting different samples and the balanced distribution of samples, we designed a new loss function. In addition, we pool the pixels in the closed area formed by the same category to simplify the number of pixel output results, and improve the segmentation speed. The test on KITTI road scene database shows that this method can effectively improve the segmentation speed while ensuring the segmentation accuracy.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
