Very Deep Convolutional Networks for Large-Scale Image Recognition
Citations
316 citations
315 citations
Cites background or methods from "Very Deep Convolutional Networks fo..."
...As argued in [19], successive convolutions by small filters equal to one convolution operation by a large filter, but effectively enhances the model’s discriminative power and reduces the number of filter parameters to learn....
[...]
...The CNN architectures designed in this paper are inspired by two previous works [19], [11], but with a number of modifications and improvements, and our designed CNN models have visible advantage in performance....
[...]
...Two CNN models named NN1 and NN2 are designed, which are closely related to the ones proposed in [19], [11], but with a number of modifications and improvements....
[...]
315 citations
315 citations
Cites background from "Very Deep Convolutional Networks fo..."
...With the recent development of convolutional neural networks (CNNs) [15, 24, 11], great progress has been made in object detection, in which object localization is generally framed as a regression problem that relocates an initial proposal to its designated target....
[...]
315 citations
Cites methods from "Very Deep Convolutional Networks fo..."
...In the case study, we consider four CNN applications (including both convolution layers and fully connected layers): 8-layer AlexNet [56], 16-layer VGG-16, 19-layer VGG-19 [85], and 152-layer ResNet-152 [41]....
[...]
...Figure 13 shows our training result for AlexNet and VGG-16 [85] on ImageNet [78]....
[...]
References
49,639 citations
21,729 citations
9,803 citations
9,775 citations
6,061 citations