News & Posts

and AI Blog

We Are Hiring! Please read the opportunities below.

Recruit Postdocs, RAs, full-time/part-time PhDs, Internships, and Research Scientists. Drop me an email pluo.lhi@gmail.com.

Introduce a new family of normalization methods in Deep Learning.

Understanding batch normalization in Deep Learning. This blog was written in Chinese.

Projects

Research Topics and Popular Benchmarks. See Publication List for more details.

*

DeepFashion2

DeepFashion second edition with a full-spectrum of fashion image analyses.

Switchable Normalization

Meta-learning to learn normalization method for each hidden layer in ConvNet.

Regularization in BN

Understanding Batch Normalization in deep learning.

Traffic Scene Segmentation

Fast scene segmentation by layer cascade deep networks.

Lane Detection

Spatial CNN for Lane Detection.

Understanding Normalization

Understanding Normalization Methods in Deep Learning.

Face Image Generation

Image Generation via GANs.

CUImage Dataset

A large-scale dataset for learning general visual representation.

Face Relationship

A large-scale face relationship dataset.

Language Guided Image Segmentation

Joint learning image and language.

WIDERFace

A large-scale dense face detection challenge.

DeepFashion

DeepFashion first edition.

Face Model Compression

An extremely fast face recognition system .

Comprehensive Car

A large-scale car re-identification benchmark.

CelebA

Face celebrity dataset for attribute recognition and GANs.

Deep Learning MRF for Image Segmentation

Deep learning for semantic image segmentation.

Pedestrian Detection

Pedestrian Detection via Rich Supervisions.

Pedestrian Parsing

A pedestrian parsing benchmark.

Talks & Workshops

Presentations, Tutorials, and Workshops

This talk was delivered in Chinese. 一个卷积层,一个归一化层,一个非线性激活函数一起构成了深度卷积神经网络 (ConvNet)的“原子”结构。

We organize a new challenge in conjunction with ECCV 2018. The challenge centers around the problem of precise localization of human …

Contact

  • pluo.lhi@gmail.com
  • Department of Computer Science, Chow Yei Ching Building, The Univeristy of Hong Kong, Pokfulam Road, Hong Kong