Seeking Visiting scholar

Seeking opportunities for Visiting scholar Ph.D

Profile

I am currently pursuing a Ph.D. at Xidian University, focusing on research areas including video image enhancement, high dynamic range video generation, and video compression.

I am looking for interdisciplinary collaborative research, including but not limited to, multi-modal perception and understanding, safe and trustworthy AI.

Research and Publications

To date, I have authored over 10 scholarly articles and actively participate as a peer reviewer for top-tier conferences such as CVPR, ICCV, ACM MM, and AAAI. My work primarily explores the intersections of image processing and data compression in digital media technologies.

Achievements

My research team secured the fourth place in the objective quality category at the CLIC 2022 video compression challenge. I am also a recipient of the National Scholarship among other awards.

🔥 News

• 2024.03: 🎉 Invited by CVPR’2025 as Reviewer.
• 2024.08: 🎉 !🔥 🤗🤗🤗 The Inference Code of PGTFormer is in huggingface [PGTFormer](Video face Restoration is released).
• 2024.08: 🎉 !🔥 5 Paper is accepted by ACM MM Workshop 2024.
• 2024.07: 🎉 !🔥 An End-to-End Real-World Camera Imaging Pipeline is accepted by ACM MM 2024.
• 2024.05: 🎉 Invited by NeurIPS as Reviewer.
• 2024.04: 🎉 Beyond Alignment: Blind Video Face Restoration via Parsing-Guided Temporal-Coherent Transformeris accepted by IJCAI 2024.
• 2024.03: 🎉 Invited by MM as Program Reviewer.

📝 Publications

ArXiv

Beyond Feature Mapping GAP: Integrating Real HDRTV Priors for Superior SDRTV-to-HDRTV Conversion

Kepeng Xu, Li Xu, Gang He, Zhiqiang Zhang, Wenxin Yu, Shihao Wang, Dajiang Zhou, Yunsong Li.

[Project]

• A more generalizable real-world SDRTV to HDRTV conversion method

ACMMM 2024

An End-to-End Real-World Camera Imaging Pipeline

Kepeng Xu, Zijia Ma, Li Xu et al.

Project

• End-to-end imaging pipeline for complex distortions

IJCAI 2024

Beyond Alignment: Blind Video Face Restoration via Parsing-Guided Temporal-Coherent Transformer

Kepeng Xu, Li Xu, Gang He et al.

Project

• Clearer and more consistent video face restoration.

Awesome-SDRTV-to-HDRTV

Awesome-SDRTV-to-HDRTV

Awesome-Video Face Restoration

Awesome-Video Face Restoration

ACM MM 2022

SDRTV-to-HDRTV via Hierarchical Dynamic Context Feature Mapping

Gang He, Kepeng Xu*, Li Xu et al.

Project

• Provides better HDR conversion quality.

• ACMMM24 EMCLR [ Neural Video Compression with Re-Parametrisation Scene Content-Adaptive Network], Kepeng Xu, Gang He† et al.
• ACMMM24 EMCLR [ Adaptive Dual-Domain Debanding: A Novel Algorithm for Image and Video Enhancement], Kepeng Xu, Gang He† et al.
• ACMMM24 Meet4MM [High-dynamic Range Video Generation Method Based On Adaptive Priors Dynamic Mapping], Gang He, SiqiWang,Kepeng Xu et al.
• ACMMM24 QoEVMA [ No-Reference Image Quality Assessment via Local and Global Multi-Scale Feature Integration], Kepeng Xu, Gang He† et al.

• ACMMM24 QoEVMA [ Banding Detection via Adaptive Global Frequency Domain Analysis], Gang He,Kepeng Xu † et al.

• Arxiv Towards Robust SDRTV-to-HDRTV via Dual Inverse Degradation Network, Kepeng Xu, Li Xu, Gang He† et al.

• CLIC 22 Hybrid video coding scheme based on VVC and spatio-temporal attention convolution neural network, Gang He, Kepeng Xu* et al.

• CVPRW 22 DNAS: A Decoupled Global Neural Architecture Search Method, Kepeng Xu*, Gang He.

• Arxiv Sdrtv-to-hdrtv conversion via spatial-temporal feature fusion, Kepeng Xu*, Li Xu et al.

• ICONIP Inpainting with Sketch Reconstruction and Comprehensive Feature Selection, S Li, L Lu, K Xu et al.

• ICONIP LPI-Net: Lightweight Inpainting Network with Pyramidal Hierarchy, S Li, L Lu, Z Li, K Xu et al.

• ICMEW Segmentation of building footprints with xception and iouloss, K Xu, Y Zhang, W Yu et al.

• ICIP Interactive separation network for image inpainting, S Li, L Lu, Z Zhang, X Cheng, K Xu et al.

• ICONIP IRSNET: An inception-ResNet feature reconstruction model for building segmentation, K Xu, L Nie, Z Zhang et al.

🎖 Honors and Awards

• 2024.09 National Scholarships for PhD Candidates.
• 2022.09 National Scholarships for Postgraduate Students.
• 2022.04 CLIC video compression PSNR ranked fourth.
• 2019.06 Learning-Based Image Inpainting Runner Award.

📖 Educations

• 2023.09 - 2027.06(now), Ph.D, Xidian University, Xi’an.
• 2020.09 - 2023.06, Master, Xidian University, Xi’an.
• 2016.09 - 2020.06, Information and Computing Science, Southwest University of Science and Technology, Mianyang.

💬 Invited Talks

• 2024.08, video face restoration IJCAI oral presentation.
• 2024.04, video enhancement technology Alipay internal talk.
• 2024.04, HDR video generation and quality assessment technology Alipay internal talk.
• 2024.04, HDR video generation and quality assessment technology Alipay internal talk.
• 2024.01, Real World HDR video generation technology Alipay internal talk.
• 2022.10, HDR video generation technology talk ACM MM.
• 2022.08, HDR video generation key technology solves MEGVII internal talk.
• 2022.05, Talk 3D Rendering Course.
• 2022.06, HDR video generation technology MEGVII internal talk.
• 2022.06, Video compression technology solution talk, CLIC2022 (CVPR).

💻 Internships

• 2022.05 - 2022.10, [MEGVII], China.