Jing Shi

Introduction

I am a fifth-year CS Ph.D student at the University of Rochester, supervised by Chenliang Xu.
I got my bechalor degree from the University of Electronic Sience and Techonology of China.
My interest spans deep learning, machine learning, and computer vision.

CV Google Scholar Linkedin Github

Connection

Office: Wegmans 2403
Email: j.shi(at)rochester(dot)edu

News

[Jul/2021] Three papers accepted by ICCV2021: two about language and scene understanding, one about language driven image editing.
[Mar/2021] Our language driven image editing paper is accepted by CVPR2021.
[Sep/2020] Our language driven image editing paper is accepted by ACCV2020 Oral.
[Feb/2019] Our weakly supervised video grounding paper is accepted by CVPR 2019.

Publications

	SpaceEdit: Learning a Unified Editing Space for Open-Domain Image Editing Jing Shi, Ning Xu, Haitian Zheng, Alex Smith, Jiebo Luo, Chenliang Xu Computer Vision and Pattern Recognition (CVPR), 2022 We introduce a pretrained pix2pix model with a unified editing space that can work for various editing-related downstream tasks.
	A Simple Baseline for Weakly-Supervised Scene Graph Generation Jing Shi, Yiwu Zhong, Ning Xu, Yin Li, Chenliang Xu International Conference on Computer Vision (ICCV), 2021 We construct a versatile baseline for Weakly-Supervised Scene Graph Generation.
	Learning to Generate Scene Graph from Natural Language Supervision Yiwu Zhong, Jing Shi, Jianwei Yang, Chenliang Xu Yin Li, International Conference on Computer Vision (ICCV), 2021 code We generate scene graph from only natural language supervision.
	Language-Guided Global Image Editing via Cross-Modal Cyclic Mechanism Wentao Jiang, Ning Xu, Jiayun Wang, Chen Gao, Jing Shi, Zhe Lin, Si Liu , International Conference on Computer Vision (ICCV), 2021 We tackle language-guided global image editing via GAN model with cross-modal consistency
	Learning by Planning: Language-Guided Global Image Editing Jing Shi, Ning Xu, Trung Bui, Franck Dernoncourt, Chenliang Xu Computer Vision and Pattern Recognition (CVPR), 2021 project page / code&dataset We contruct a new framework for language-guided global image editing.
	Cubic Spline Smoothing Compensation for Irregularly Sampled Sequences Jing Shi, Jing Bi, Yingru Liu, Chenliang Xu Arxiv Preprint We introduce the cubic spline smoothing compensation upon ODE-RNN model for irregularly observed time series.
	Learning Continuous-Time Dynamics by Stochastic Differential Networks Yingru Liu, Yucheng Xing, Xuewen Yang, Xin Wang, Jing Shi, Di Jin, Zhaoyue Chen Arxiv Preprint
	A Benchmark and Baseline for Language-Driven Image Editing Jing Shi, Ning Xu, Trung Bui, Franck Dernoncourt, Zheng Wen, Chenliang Xu Asian Conference on Computer Vision (ACCV), 2020 (Oral) code / dataset We introduce the GIER dataset and a pipeline for language-driven image editing.
	GAN-EM: GAN Based EM Learning Framework Wentian Zhao, Shaojie Wang, Zhihuai Xie, Jing Shi, Chenliang Xu International Joint Conference on Artificial Intelligence (IJCAI), 2019 code
	Not All Frames Are Equal: Weakly-Supervised Video Groundingwith Contextual Similarity and Visual Clustering Losses Jing Shi, Jia Xu, Boqing Gong, Chenliang Xu Computer Vision and Pattern Recognition (CVPR), 2019 project page/ code We introduce contextual smilarity loss and visual clustering loss for weakly supervised video grounding.
	Audio-Visual Event Localization in Unconstrained Videos Yapeng Tian, Jing Shi, Bochen Li, Zhiyao Duan, Chenliang Xu European Conference on Computer Vision (ECCV), 2018 project page/ code We introduce audio-visual event localization to understand the interaction of vision and audio.
	Boundary Vibration Control of Variable Length Crane Systems in Two Dimensional Space with Output Constraints Xiuyu He, Wei He, Jing Shi, Changyin Sun IEEE/ASME Transactions on Mechatronics (TMech), 2017 We demonstrate the lyapunov stability of the boundary control for the moving crane system in 2-D space.
	Impact Damage Detection and Characterization using Eddy Current Pulsed Thermography Yizhe Wang, Han Ke, Jing Shi, Bing Gao, Guiyun Tian IEEE Far East NDT New Technology & Application Forum (FENDT), 2016 We use eddy current pulse thermography to detect impact damages of torque arm in aircraft brake system.

Experience

Project

Differential Network for Video Object Detection
Jing Shi, Chenliang Xu
CSC 577 Advanced Topics in Computer Vision, 2017
code

For flow based video obejct detection, we propose the differential network to select the key frame. Then we adapt the sequential NMS to an incremental form so that it is both time-efficient and accurate.

Internship

Tencent AI Lab , Shenzhen, May - Aug 2018

Collaborator: Jia Xu, Boqing Gong
Project: Weakly supervised video grounding.

Adobe Research , San Jose, May - Sep 2019

Collaborator: Ning Xu,
Project: Language Driven Image Editing.

Adobe Research , San Jose, May - Sep 2021

Collaborator: Ning Xu,
Project: Unified Model for Various Image Editing Tasks.

Course

CSC 400 Problem Seminar	CSC 440 Data Mining
CSC 454 Program Language Design and Implementation	CSC 577 Advanced Topics in Computer Vision
CSC 446 Machine Learning	CSC 484 Advanced Algorithms
CSC 455 Software Analysis and Improvement	CSC 453 Dynamic Language and Software Development
CSC 480 Computing Model and Limitation	CSC 249/449 Machine Vision [Teaching Assistant]