李成华,中国科学院自动化研究所副研究员,2018年在中国科学院自动化研究所博士毕业,并留所工作至今。目前主要从事深度学习、图像与视频内容分析方面研究,在图像视频增强领域获得国际著名竞赛NTIRE、AIM冠亚季军10余项,发表学术论文10余篇,申请发明专利20余项。现为中国图象图形学会会员、CSIG视觉大数据专委会委员,荣获南京市省第十五批科技镇长团优秀团员。
图像增强
北京三星通信技术研究有限公司, 创新探索项目, SRC-Beijing-DVL-2021-00237,
软硬件友好的图像超分算法研究合作项目, 2021-10至2022-10,45万元,结题,主持
(1) Qian, Ming; Congyu Qiao; Jiamin, Lin; Zhenyu, Guo; Chenghua, Li; Cong, Leng; Jian, Cheng ; BGGAN: Bokeh-Glass Generative Adversarial Network for Rendering Realistic Bokeh, Computer Vision -- ECCV 2020 Workshops, AIM 2020 on Rendering Realistic Bokeh Track1: CPU和Track 2: smartphone GPU双赛道冠军(参赛队伍指导老师)., Glasgow, United Kingdom, 2020-8-23至2020-8-28 (会议论文) ( 本人标注: 共同通讯作者 )
(2) Zhu, Yu; Zhenyu, Guo; Tian, Liang; Xiangyu, He; Chenghua, Li; Cong, Leng; Bo, Jiang; Yifan, Zhang; Jian, Cheng ; EEDNet: Enhanced Encoder-Decoder Network for AutoISP, AIM 2020 on Learned ISP Track1: Fidelity冠军(参赛队伍指导老师)., Glasgow, United Kingdom, 2020-8-23至2020- 8-28 (会议论文) ( 本人标注: 共同通讯作者 )
(3) Shuai, Liu; Chenghua, Li; Nan, Nan; Ziyao, Zong; Ruixia, Song ; MMDM: Multi-frame and Multi-scale for Image Demoiréing, 2020 CVPR Workshops, NTIRE 2020 Challenge on Real Image Demoireing Track 1: Single Image季军、Track 2: Burst亚军;NTIRE 2020 Challenge on Real Image Denoising: Track 2 sRGB 亚军(参赛队伍指导老师)., Seattle, WA, USA, 2020-6-14至2020-6-19 (会议论文) ( 本人标注: 唯一通讯作者 )
(4) Xiaoyi Dong; Yu Zhu; Chenghua Li; Peisong Wang; Jian Cheng ; RISPNet: A Network for Reversed Image Signal Processing, ECCV 2022: Computer Vision – ECCV 2022 Workshops, AIM 2022 Challenge on Reversed ISP(Track 1 S7,Track 2 P20)双赛道亚军(参赛队伍指导老师)., Tel Aviv, Israel, 2022-8-23至2022-8-27 (会议论文) ( 本人标注: 共同通讯作者 )
(5) Xiaoran Qin; Yu Zhu; Chenghua Li; Peisong Wang; Jian Cheng ; CIDBNet: A Consecutively- Interactive Dual-Branch Network for JPEG Compressed Image Super-Resolution, ECCV 2022: Computer Vision – ECCV 2022 Workshops, ECCV 2022 Workshops, AIM 2022 Challenge on Super-Resolution on Compressed Image第三名(参赛队伍指导老师)., Tel Aviv, Israel, 2022-8-23至2022-8-27 (会议论 文) ( 本人标注: 共同通讯作者 )
1.AIM 2022 Challenge on Reversed ISP(Track 1 S7,Track 2 P20)双赛道亚军;
2.AIM 2022 Challenge on Defilter(滤镜去除)赛道亚军;
3.AIM 2022 Challenge on Super-Resolution of Compressed Image季军;
4.AIM 2020 Challenge on Learned ISP Track 1: Fidelity冠军;
5.AIM 2020 Challenge on Rendering Realistic Bokeh Track1:CPU和Track 2: smartphone GPU 双赛道冠军;
6.AIM 2020 Challenge on Scene Relighting and Illumination Estimation: Track 2 Illumination Settings Estimation冠军;
7.AIM 2020 Challenge on Efficient Super-Resolution 亚军;
8.NTIRE 2020 Challenge on Real Image Denoising: Track 2 sRGB 亚军;
9.NTIRE 2020 Challenge on Real Image Demoireing Track 1: Single Image季军、Track 2: Burst亚军;
10.AIM 2019 Challenge on Constrained Super-Resolution Track 1: Parameters亚军及Track2: Inference季军;
11.2016.04获得MSR图像精细分类竞赛冠军;
(1) 冷聪; 李成华 ; 压缩图像超分方法、图像压缩方法及系统, 2023-03-24, 中国, ZL 2022 1 1276433.0 (专利)
(2) 冷聪; 李成华 ; 一种基于图像去噪思想的图像压缩及滤镜去除方法及系统, 2023-03-24, 中国, ZL 2022 1 1270285.1 (专利)
(3) 冷聪; 李成华; 朱宇; 程健 ; 一种基于卷积神经网络的图像夜景处理方法及运行该方法的计算模块与 可读存储介质, 2023-07-07, 中国, ZL 2020 1 1049682.7 (专利)
(4) 冷聪; 李成华; 于浩东; 周波; 程健 ; 一种基于滤波器融合的超分辨率图像重建方法及系统, 2023- 08-08, 中国, ZL 2020 1 1070223.7 (专利)