Dataset and Evaluation of Interest Point Detectors for Visual Tracking
نویسندگان
چکیده
In this report, we present an extensive dataset of 96 video streams with ground truth. It includes various geometric changes, lighting conditions, and levels of motion blur, and can be used as testbed for a variety of algorithms in the context of visual tracking. We then use this dataset for a detailed quantitative evaluation of popular interest point detectors, which, in contrast to existing evaluations, is geared towards visual tracking in all relevant factors of the evaluation design.
منابع مشابه
In-Depth Evaluation of Popular Interest Point Detectors on Video Streams
We present an in-depth evaluation of popular interest point detectors, which, in contrast to existing evaluations, is targeted towards the application in visual tracking and augmented reality. In particular, candidate algorithms, testbed, and performance criteria are chosen with respect to the application of visual tracking. We evaluate the impact of individual algorithm parameters and present ...
متن کاملA thermal infrared dataset for evaluation of short-term tracking methods
During recent years, thermal cameras have decreased in both size and cost while improving image quality. The area of use for such cameras has expanded with many exciting applications, many of which require tracking of objects. While being subject to extensive research in the visual domain, tracking in thermal imagery has historically been of interest mainly for military purposes. The available ...
متن کاملEvaluation of interest point detectors for Visual SLAM
In this paper we present several interest points detectors and we analyze their suitability when used as landmark extractors for vision-based simultaneous localization and mapping. For this purpose, we evaluate the detectors according to their repeatability under changes in viewpoint and scale. These are the desired requirements for visual landmarks. Several experiments were carried out using s...
متن کاملA Comparison of Feature Detectors with Passive and Task-Based Visual Saliency
This paper investigates the coincidence between six interest point detection methods (SIFT, MSER, Harris-Laplace, SURF, FAST & Kadir-Brady Saliency) with two robust “bottom-up” models of visual saliency (Itti and Harel) as well as “task” salient surfaces derived from observer eye-tracking data. Comprehensive statistics for all detectors vs. saliency models are presented in the presence and abse...
متن کاملEvaluation of interest point detectors for image information extraction
Interest points extraction and matching is a common task in many computer vision based application, which are used in different domains, such as 3D reconstruction, object recognition, or tracking. We present an evaluation of current state of the art about interest point extraction algorithms to measure several parameters, such as detection quality, invariance to rotation and scale transformatio...
متن کامل