Google Net
Google Net
• Example
• 5×5×3 (F: 1×1 depth 3) 5×5
• 5×5×3 (F: 3×3 depth 3) 3×3
• 5×5×3 (F: 5×5 depth 3) 1×1
Applying various kernels on the given input image
• Example
• 5×5×3 (F: 1×1 depth 3) 5×5
• 5×5×3 (F: 3×3 depth 3, p=1) 5×5
• 5×5×3 (F: 5×5 depth 3, p=2) 5×5
Importance of 1×1 convolution
https://medium.com/coinmonks/paper-review-of-googlenet-inception-v1-winner-of-ilsvlc-2014-image-classification-c2b3565a64e7
Average Pooling
https://medium.com/coinmonks/paper-review-of-googlenet-inception-v1-winner-of-ilsvlc-2014-image-classification-c2b3565a64e7
C. Szegedy et al., "Going deeper with convolutions," 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 1-9, doi: 10.1109/CVPR.2015.7298594.
Inception module
• Input- 28*28*192
• 64 1*1 28*28*64
• 96 1*1 28*28*96 (3*3,128) 28*28*128
• 16 1*1 28*28*16 (5*5,32) 28*28*32
• Max pool(3*3) 26*26*192 28*28*32
• 28*28*256
References
1. C. Szegedy et al., "Going deeper with
convolutions," 2015 IEEE Conference on
Computer Vision and Pattern Recognition
(CVPR), 2015, pp. 1-9, doi:
10.1109/CVPR.2015.7298594.
2. https://medium.com/coinmonks/paper-review-of-
googlenet-inception-v1-winner-of-ilsvlc-2014-
image-classification-c2b3565a64e7
3. Deep Learning-Part1: NPTEL Course
https://nptel.ac.in/courses/106/106/106106184/#