Very Deep Convolutional Networks for Large-Scale Image Recognition
Citations
293 citations
Cites background from "Very Deep Convolutional Networks fo..."
...Driven by the advance of fully convolutional networks (FCNs) [29], human parsing, or semantic part segmentation has recently witnessed great progress thanks to deeply learned features [37,14], large-scale annotations [24,11], and advanced reasoning over graphical models [45,3]....
[...]
293 citations
Cites background from "Very Deep Convolutional Networks fo..."
...The perceptual loss is defined as the Euclidean distance between the feature representations of a reconstructed image G(R) and the reference image I from the ReLU activation layers of the pre-trained 16 layer VGG network [55]....
[...]
292 citations
292 citations
Cites background from "Very Deep Convolutional Networks fo..."
...…and classification problems has, in large part, been due to the use of convolutional network architectures that reduce dramatically the dimensionality of the solution space by enforcing highly symmetric patterns in the weights to be learned (LeCun et al. 1998, 2015; Simonyan & Zisserman 2014)....
[...]
292 citations
Cites methods from "Very Deep Convolutional Networks fo..."
...The image is processed by VGG network [23] and the activations in the last pooling layer are extracted as features for image regions....
[...]
...In our model, we select to use VGG network [23] as the frame-level appearance feature extractor because it is widely used and shows promising results in the literature....
[...]
References
49,639 citations
21,729 citations
9,803 citations
9,775 citations
6,061 citations