An Efficient 3D Model Retrieval Method Based on Convolutional Neural Network

Table 1

The structure of ResNet50.

Layer name	Output size	50-layer
Conv1	112  112	7  7, 64, stride 2
	56  56	3  3 max pool, stride 2
Conv2_x	56  56
Conv3_x	28  28
Conv4_x	14 × 14
Conv5_x	7  7
	1  1	Average pool, 1000-d fc, softmax
FLOPs		3.8  10^9