Profile cover photo
Profile photo
Multimedia Laboratory, CUHK
450 followers
450 followers
About
Posts

Post has attachment
Ping Luo, Jiamin Ren, and Zhanglin Peng, Differentiable Learning-to-Normalize via Switchable Normalization, arXiv:1806.10779

https://arxiv.org/abs/1806.10779

We address a learning-to-normalize problem by proposing Switchable Normalization (SN), which learns to select different operations for different normalization layers of a deep neural network (DNN). SN adapts to various network architectures and tasks. It is robust to a wide range of batch sizes, maintaining high performance when small minibatch is presented (e.g. 2 images/GPU). Extensive evaluations show that SN outperforms its counterparts on image classification in ImageNet, object detection and segmentation in COCO, artistic image stylization, and neural architecture search. The code of SN will be made available in https://github.com/switchablenorms/

Post has attachment
Depth Map Super-Resolution by Deep Multi-Scale Guidance, ECCV 2016

Code is released!
Project Page: http://mmlab.ie.cuhk.edu.hk/projects/guidance_SR_depth.html

Post has shared content
PolyNet: A Pursuit of Structural Diversity in Very Deep Networks

https://arxiv.org/abs/1611.05725

We explore structural diversity in designing deep networks, a new dimension beyond just depth and width. Specifically, we present a new family of modules, namely the PolyInception, which can be flexibly inserted in isolation or in a composition as replacements of different parts of a network. Choosing PolyInception modules with the guidance of architectural efficiency can improve the expressive power while preserving comparable computational cost. A benchmark on the ILSVRC 2012 validation set demonstrates substantial improvements over the state-of-the-art. Compared to Inception-ResNet-v2, it reduces the top-5 error on single crops from 4.9% to 4.25%, and that on multi-crops from 3.7% to 3.45%.
Photo
Add a comment...

Post has attachment
PolyNet: A Pursuit of Structural Diversity in Very Deep Networks

https://arxiv.org/abs/1611.05725

We explore structural diversity in designing deep networks, a new dimension beyond just depth and width. Specifically, we present a new family of modules, namely the PolyInception, which can be flexibly inserted in isolation or in a composition as replacements of different parts of a network. Choosing PolyInception modules with the guidance of architectural efficiency can improve the expressive power while preserving comparable computational cost. A benchmark on the ILSVRC 2012 validation set demonstrates substantial improvements over the state-of-the-art. Compared to Inception-ResNet-v2, it reduces the top-5 error on single crops from 4.9% to 4.25%, and that on multi-crops from 3.7% to 3.45%.
Photo

Post has attachment
Chao Dong, Chen Change Loy, Xiaoou Tang, Accelerating the Super-Resolution Convolutional Neural Network, ECCV 2016

Project Page and Codes
http://mmlab.ie.cuhk.edu.hk/projects/FSRCNN.html

Add a comment...

Post has attachment
New state-of-the-art action recognition with a framework called temporal segment networks.

Paper:
L. Wang, Y. Xiong, Z. Wang, Y. Qiao, D. Lin, X. Tang, L. Van Gool,
Temporal Segment Networks: Towards Good Practices for Deep Action Recognition, ECCV 2016

Arxiv Preprint:
http://arxiv.org/abs/1608.00859v1

Code & Models
https://github.com/yjxiong/temporal-segment-networks
Photo
Add a comment...

Post has attachment
Fashion landmark is a more discriminative representation than human joints and bounding boxes
to understand fashion images. A large-scale fashion landmark detection benchmark will be released together with the DeepFashion database.

Paper:
Ziwei Liu, Sijie Yan, Ping Luo, Xiaogang Wang, and Xiaoou Tang, Fashion Landmark Detection in the Wild, ECCV 2016

DeepFashion:
http://mmlab.ie.cuhk.edu.hk/projects/DeepFashion.html
Add a comment...

Post has attachment
S. Zhu, S. Liu, C. C. Loy, X. Tang, Deep Cascaded Bi-Network for Face Hallucination, ECCV 2016

Paper:
http://personal.ie.cuhk.edu.hk/~ccloy/files/eccv_2016_hallucination.pdf

Technical Report:
http://arxiv.org/abs/1607.05046

Code:
https://github.com/zhusz/ECCV16-CBN
Photo
Add a comment...

Post has attachment
Learning Deep Representation for Imbalanced Classification

Project Page and Code:
http://mmlab.ie.cuhk.edu.hk/projects/LMLE.html

Add a comment...
Wait while more posts are being loaded