Jingxiang Sun (孙景翔)

Email: sunjingxiang_stark[at]126.com       Google Scholar      Github

I am a first-year PhD student of Department of Automation, Tsinghua University in fall, 2022 and under the supervision of Prof. Yebin Liu. Prior to Tsinghua University I obtained M.S. from Department of ECE, University of Illinois at Urbana-Champaign and B.S. from Beijing University of Technology. My research focuses on neural rendering, digital avatar and image generation.

I am very fortunate to have spent summer 2021 at Tencent AI Lab with Xuan Wang and Jue Wang; fall 2020 at State Key Laboratory of CAD&CG at Zhejiang University with Xiaowei Zhou.

We propose a novel 3D GAN framework for unsupervised learning of generative, high-quality and 3D-consistent facial avatars from unstructured 2D images. To achieve both deformation accuracy and topological flexibility, we present a 3D representation called Generative Texture-Rasterized Tri-planes.

We propose an efficient method to construct the personalized generative prior based on a small set of facial images of a given individual. After learning, it allows for photo-realistic rendering with novel views and the face reenactment can be realized by performing navigation in the latent space. Our proposed method is applicable for different driven signals, including RGB images, 3DMM coefficients, and audios.

We propose DiffuStereo, a novel system using only sparse cameras (8 in this work) for high-quality 3D human reconstruction. At its core is a novel diffusion-based stereo module, which introduces diffusion models, a type of powerful generative models, into the iterative stereo matching network.

We propose a high-resolution 3D-aware generative model that not only enables local control of the facial shape and texture, but also supports real-time, interactive editing.

We propose FENeRF, a 3D-aware generator that can produce view-consistent and locally-editable portrait images. Our method uses two decoupled latent codes to generate corresponding facial semantics and texture in a spatial aligned 3D volume with shared geometry. We also reveal that joint learning semantics and texture helps to generate finer geometry.

We propose a novel optimization-based framework and experimentally demonstrate its ability to recover much more precise and detailed motion from multiple videos, compared against monocular pose estimation methods.

We propose a general and practical evaluation framework for analysing various widely used prediction models (i.e. delay, k- nearest-neighbor, kernel regression, additive model, and recur- rent neural network using long short term memory) for bus arrival time.

The website template was adapted from Yu Deng.