Qiqi Hou

I am a machine learning researcher at Qualcomm AI Research. Much of my current research is about the applications of deep learning, such as data compression, Monte Carlo rendering, perceptual quality metrics, frame interpolation, matting, face landmark detection and etc. Before that, I received my Ph.D. and M.E. degrees from Portland State University and Xi'an Jiaotong University in 2023 and 2017, respectively, where I was advised by Feng Liu and Jinjun Wang, respectively.

Email  /  CV  /  Google Scholar  /  Github  / 

profile photo
Selected Research
Neural Graphics Texture Compression Supporting Random Access
Farzad Farhadzadeh, Qiqi Hou, Hoang Le, Amir Said, Randall Rauwendaal, Alex Bourd, Fatih Porikli
ECCV, 2024  
arxiv

We propose an asymmetric auto-encoder to compress the graphics texture.

Low-Latency Neural Stereo Streaming
Qiqi Hou, Farzad Farhadzadeh, Amir Said, Guillaume Sautiere, Hoang Le
CVPR, 2024  
arxiv / poster / video-demo

We propose a parallel codec for fast and efficient low-latency stereo video compression.

Auxiliary Features-Guided Super Resolution for Monte Carlo Rendering
Qiqi Hou, Feng Liu
Computer Graphics Forum, 2023  
arxiv / code (coming) / model (coming) / data (coming) / 3min-demo / interactive-viewer

We accelerate Monte-Carlo rendering via super-resoultion guided by high-resolution fast-to-compute auxiliary features.

A Perceptual Quality Metric for Video Frame Interpolation
Qiqi Hou, Abhijay Ghildyal, Feng Liu
ECCV, 2022  
arxiv / code / model / data / result / poster / 3min-demo

We propose a perceptual quality metric dedicated to video frame interpolation.

Fast Monte Carlo Rendering via Multi-Resolution Sampling
Qiqi Hou*, Zhan Li*, Carl S Marshall, Selvakumar Panneer, Feng Liu
Graphic Interface, 2021  
arxiv / code / model / data / result / 5min-talk

We accelerate the Monte-Carlo rendering via multiple resolution sampling. We also introduce a large scale ray-tracing dataset - BCR dataset.

Context-Aware Image Matting for Simultaneous Foreground and Alpha Estimation
Qiqi Hou, Feng Liu
ICCV, 2019  
arxiv / code / model / result / poster

We can estimate the alpha matte and foreground simultaneously.

Face Alignment Recurrent Network
Qiqi Hou, Jinjun Wang, Ruibin Bai, Sanping Zhou, Yihong Gong
Pattern Recognition, 2018  
paper / code

A face align recurent network to help avoid over-strong early stage regressors and over-weak later stage regressors.

Awards
cs188 El-Mansy Family Fellowship

Richard Kieburtz Memorial Graduate Fellowship

Last updated: Oct, 2022. Template from here.