TY - GEN
T1 - A biologically inspired deep CNN model
AU - Zhang, Shizhou
AU - Gong, Yihong
AU - Wang, Jinjun
AU - Zheng, Nanning
N1 - Publisher Copyright:
© Springer International Publishing AG 2016.
PY - 2016
Y1 - 2016
N2 - Recently, the Deep Convolutional Neural Networks (DCNN) have achieved state-of-the-art performances with many tasks in image and video analysis. However, it is a very challenging problem to devise a good DCNN model as there are so many choices to be made by a network designer, including the depth, the number of feature maps, interconnection patterns, window sizes for convolution and pooling layers, etc. These choices constitute a huge search space that makes it impractical to discover an optimal network structure with any systematic approaches. In this paper, we strive to develop a good DCNN model by borrowing biological guidance from the human visual cortex. By making an analogy between the proposed DCNN model and the human visual cortex, many critical design choices of the proposed model can be determined with some simple calculations. Comprehensive experimental evaluations demonstrate that the proposed DCNN model achieves state-of the- art performances on four widely used benchmark datasets: CIFAR- 10, CIFAR-100, SVHN and MNIST.
AB - Recently, the Deep Convolutional Neural Networks (DCNN) have achieved state-of-the-art performances with many tasks in image and video analysis. However, it is a very challenging problem to devise a good DCNN model as there are so many choices to be made by a network designer, including the depth, the number of feature maps, interconnection patterns, window sizes for convolution and pooling layers, etc. These choices constitute a huge search space that makes it impractical to discover an optimal network structure with any systematic approaches. In this paper, we strive to develop a good DCNN model by borrowing biological guidance from the human visual cortex. By making an analogy between the proposed DCNN model and the human visual cortex, many critical design choices of the proposed model can be determined with some simple calculations. Comprehensive experimental evaluations demonstrate that the proposed DCNN model achieves state-of the- art performances on four widely used benchmark datasets: CIFAR- 10, CIFAR-100, SVHN and MNIST.
UR - https://www.scopus.com/pages/publications/85007210886
U2 - 10.1007/978-3-319-48890-5_53
DO - 10.1007/978-3-319-48890-5_53
M3 - 会议稿件
AN - SCOPUS:85007210886
SN - 9783319488899
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 540
EP - 549
BT - Advances in Multimedia Information Processing – 17th Pacific-Rim Conference on Multimedia, PCM 2016, Proceedings
A2 - Chen, Enqing
A2 - Tie, Yun
A2 - Gong, Yihong
PB - Springer Verlag
T2 - 17th Pacific-Rim Conference on Multimedia, PCM 2016
Y2 - 15 September 2016 through 16 September 2016
ER -