The implementation of an object detection algorithm on the FT processor
Keyword(s):
With the continuous development of automatic drive and neural networks, it is possible to use neural network algorithm to carry out object detection in unmanned driving. Usually, the computation of neural network algorithm is huge. How to efficiently compute the algorithm and meet the real-time requirement is a challenge. In this paper, a sparse neural network algorithm is proposed, which can improve the utilization rate of processors. The object detection algorithm YOLO is implemented on the processor. Its performance is equivalent to the current best processor performance.
2011 ◽
Vol 90-93
◽
pp. 2173-2177
1995 ◽
Vol 365
(1)
◽
pp. 198-202
◽
Keyword(s):
2016 ◽
Vol 7
(3)
◽
pp. 292
◽
2014 ◽
Vol 513-517
◽
pp. 687-690
◽
Keyword(s):
Keyword(s):