Huikai Wu

About Me

I’m currently a PhD candidate student in Computer Vision at NLPR, CASIA. My research intersts include image processing, computer vision, deep learning and reinforcement learning. I’m focusing on dense image prediction, neural architecture search and model acceleration.

[Resume: Industry/Academia] [GitHub] [Google Scholar] [Mail]

News

• 2019.07: One paper is accepted by ICCV 2019.
• 2019.07: One paper is accepted by ACMMM 2019, oral.
• 2019.05: One paper is accepted by TIP.
• 2019.04: Survey for Neural Architecture Search—AWESOME NAS.
• 2019.04: Release a website for DEMOS of the past research.
• 2018.02: Two papers are accepted by CVPR 2018.

Research

2019

• FastFCN: Rethinking Dilated Convolution in the Backbone for Semantic Segmentation
  Huikai Wu, Junge Zhang, Kaiqi Huang, Kongming Liang, Yizhou Yu
  arXiv preprint arXiv:1903.11816
  [Project] [Paper] [arXiv] [GitHub]

2018

• SparseMask: Differentiable Connectivity Learning for Dense Image Prediction
  Huikai Wu, Junge Zhang, Kaiqi Huang
  IEEE International Conference on Computer Vision (ICCV), 2019
  [Project] [Paper] [arXiv] [GitHub]
• Fast A3RL: Aesthetics-Aware Adversarial Reinforcement Learning for Image Cropping
  Debang Li, Huikai Wu, Junge Zhang, Kaiqi Huang
  IEEE Transactions on Image Processing ( Volume: 28 , Issue: 10 , Oct. 2019 )
  [Paper]

2017

• Fast End-to-End Trainable Guided Filter
  Huikai Wu, Shuai Zheng, Junge Zhang, Kaiqi Huang
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018
  [Project] [Paper] [arXiv] [Demo] [GitHub]
• A2-RL: Aesthetics Aware Reinforcement Learning for Image Cropping
  Debang Li, Huikai Wu, Junge Zhang, Kaiqi Huang
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018
  [Project] [arXiv] [Demo] [GitHub]
• MSC: A Dataset for Macro-Management in StarCraft II
  Huikai Wu, Junge Zhang, Kaiqi Huang
  arXiv preprint arXiv:1710.03131
  [arXiv] [GitHub]
• GP-GAN: Towards Realistic High-Resolution Image Blending
  Huikai Wu, Shuai Zheng, Junge Zhang, Kaiqi Huang
  ACM International Conference on Multimedia (ACMMM), 2019, oral
  [Project] [arXiv] [GitHub]

Project

2019

• DeepJS
  Online demos for my research on image processing and computer vision based on deep learning.

2018

• FaceSwap
  Swap faces between two photos with Python 3, OpenCV and dlib.

2017

• a3c-deepmind-pytorch & pytorch-dqn-universe
  A3C and Deep Q-Learning in PyTorch.
• torchcraft-py3 & gym-starcraft-py3
  TorchCraft and gym-starcraft in Python 3.
• chainer-realism-cnn
  Chainer implementation of realismCNN proposed in “Learning a Discriminative Model for the Perception of Realism in Composite Images”.
• chainer-neural-style
  Chainer implementation of “A Neural Algorithm of Artistic Style” and “Perceptual Losses for Real-Time Style Transfer and Super-Resolution”.
• chainer-pix2pix
  Chainer implementation of “Image-to-Image Translation Using Conditional Adversarial Networks”.

2016

• ProximalGradientOptimization
  Python source code & example for using proximal gradient optimization.
• RayTracing [Video]
  RayTracing framework using C++.
• SIFT & HOG_CPP
  SIFT and HOG in C++.

2015

• SophiaGL & OpenGLWrapper
  SophiaGL provides a set of easy-to-use APIs for OpenGL 3 and above, while OpenGLWrapper contains wrappers for OpenGL to provide easy_use API for rendering 3D world.

Experience

Education

• Ph.D. in Computer Vision, Institute of Automation, Chinese Academic of Sciences
  Sep. 2015 - Present
• B.Eng. in Software Engineering, Nanjing University
  Sep. 2011 - July 2015

Internship

• International Intern in Preferred Networks (Tokyo, Japan)
  July. 2018 - Oct. 2018
• Researcher in Palmwin Information Technology (Nanjing, China)
  Aug. 2015 - Oct. 2015

Reviewer

• CVPR 2019-2020, ICCV 2019, ECCV 2020, AAAI 2020
• TPAMI, TMI

Awards

• 2017: 4th Place, AIIDE StarCraft AI Competition 2017
• 2016: 1st Place, CCF Big Data Competition: Movie Box Prediction.