Title :
Performance Evaluation of Neuromorphic-Vision Object Recognition Algorithms
Author :
Kasturi, R. ; Goldgof, D.B. ; Ekambaram, R. ; Sharma, R. ; Pratt, G. ; Anderson, M. ; Peot, M. ; Aguilar, M. ; Krotkov, E. ; Hackett, D.D. ; Khosla, D. ; Yang Chen ; Kyungnam Kim ; Yang Ran ; Qinfen Zheng ; Elazary, L. ; Voorhies, R.C. ; Parks, D.F. ; Itt
Author_Institution :
Dept. of Comput. Sci. & Eng., Univ. of South Florida, Tampa, FL, USA
Abstract :
The U.S. Defense Advanced Research Projects Agency´s (DARPA) Neovision2 program aims to develop artificial vision systems based on the design principles employed by mammalian vision systems. Three such algorithms are briefly described in this paper. These neuromorphic-vision systems´ performance in detecting objects in video was measured using a set of annotated clips. This paper describes the results of these evaluations including the data domains, metrics, methodologies, performance over a range of operating points and a comparison with computer vision based baseline algorithms.
Keywords :
computer vision; military computing; object recognition; U.S. Defense Advanced Research Projects Agency Neovision2 program; annotated clips; artificial vision system; computer vision; mammalian vision systems; neuromorphic-vision object recognition algorithm; Algorithm design and analysis; Educational institutions; Feature extraction; Neuromorphics; Object recognition; Retina; Visualization; Neuromorphic vision; object detection; performance evaluation; recognition; video analysis;
Conference_Titel :
Pattern Recognition (ICPR), 2014 22nd International Conference on
Conference_Location :
Stockholm
DOI :
10.1109/ICPR.2014.416
