no code implementations • CVPR 2018 • Falong Shen, Shuicheng Yan, Gang Zeng
Recent works on style transfer typically need to train image transformation networks for every new style, and the style is encoded in the network parameters by enormous iterations of stochastic gradient descent, which lacks the generalization ability to new style in the inference stage.
1 code implementation • 13 Sep 2017 • Falong Shen, Shuicheng Yan, Gang Zeng
Recent works on style transfer typically need to train image transformation networks for every new style, and the style is encoded in the network parameters by enormous iterations of stochastic gradient descent.
1 code implementation • CVPR 2017 • Falong Shen, Rui Gan, Shuicheng Yan, Gang Zeng
The proposed joint model also employs a guidance CRF to further enhance the segmentation performance.
no code implementations • 28 May 2016 • Falong Shen, Gang Zeng
The weighted residual network is able to learn to combine residuals from different layers effectively and efficiently.
no code implementations • 13 May 2016 • Falong Shen, Gang Zeng
This paper describes a fast and accurate semantic image segmentation approach that encodes not only the discriminative features from deep neural networks, but also the high-order context compatibility among adjacent objects as well as low level image features.