Abstract
An advanced and user-friendly tool for fast labeling of moving objects captured with surveillance sensors is proposed, which is available to the public. This tool allows the creation of three kinds of labels: moving objects, shadows and occlusions. These labels are created at both the pixel level and object level, which makes them suitable to assess the quality of both moving object detection strategies and tracking algorithms. The labeling can be performed easily and quickly thanks to a very friendly graphical user interface that allows one to automatize many common operations. This interface also includes some semiautomatic advanced tools that simplify the labeling tasks and drastically reduce the time required to obtain high-quality results.
Highlights
In spite of the interest of many researchers in the development of new and complete databases to assess the quality of their moving object detection and tracking algorithms, to the best of our knowledge, there are no public tools designed to perform the labeling of moving objects in video sequences
We propose a new and user-friendly labeling tool named TSLAB (Tool for Semiautomatic Labeling), which is available to the public on a website [18]
With the aim of being able to provide ground-truth data suitable for strategies that try to detect and/or delete shadows cast by moving objects or tracking methods that require knowing the visible regions of moving objects, and their occluded ones, TSLAB allows the labeling of three kinds of data: moving objects, shadows cast by moving objects and occluded regions of moving objects
Summary
The number of computer vision applications has exponentially grown during the last few years [1,2], mainly due to the huge amount of consumer electronic devices with optical sensors that claim them, such as smart-phones, tablets, video game consoles or intelligent vehicles [3]. The only existing public tool that could be used for this purpose is LabelMe Video [17], which is an online tool that allows one to annotate forms and names of any kind of object in an image, as well as events This tool is not focused on labeling moving objects at the pixel level and along all of the images in a video sequence. We propose a new and user-friendly labeling tool named TSLAB (Tool for Semiautomatic Labeling), which is available to the public on a website [18] This tool allows one to efficiently create ground-truth data of video sequences at the pixel level and at the object level.
Sign-in/Register to view highlights & summary 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.
