简介
曾联合新加坡国立大学LV组参加PASCAL VOC(视觉物体分类)连续三年获奖,在美国国家技术标准局(NIST)主办的非受限条件下人脸识别竞赛IJB-A获得了人脸验证(verification)与人脸辨认(identification)的双项冠军,在微软百万名人识别竞赛MS-Cele-1M人脸挑战赛获双项冠军。
毕业于西安电子科技大学电子工程系,获电子信息工程硕士学位。
人物履历
2019年,加入澎思科技,担任公司首席科学家及澎思新加坡研究院院长;
2016年,担任新加坡松下研究院副院长;
1992年,加入松下,在新加坡松下研究院担任高级主任工程师和研发主管;
1984年,毕业于西安电子科技大学,获电子工程系电子信息工程硕士学位。
研究领域
主要的研究领域是AI和深度学习,拥有计算机视觉的全栈技术能力,技术面横跨多个应用领域。技术包括人脸识别,目标跟踪、行人检测、行人再识别(RE-ID),目标检测、图像分割、图像分类,交通异常检测、自动驾驶,自动情感识别,智能机器人等,致力于监控与安全、智慧城市、自动驾驶、智能机器人以及AI工厂自动化解决方案等领域的落地。
团队获奖记录
1.2018年12月,获NIPS的自动驾驶比赛AI-DO(AI Driverless Car Olympic)大赛冠军。
2.2018年10月,在机器人学术会议IROS上获得“移动操作机器人竞赛”冠军。
3.2018年7月,获得CVPR智慧城市挑战赛“无监督交通异常检测”任务冠军。
4.2017年3月和2018年2月,参加美国国家技术标准局(NIST)主办的非受限条件下人脸识别竞赛IJB-A,获得了人脸验证(verification)与人脸辨认(identification)的双项冠军。
5.2017年7月,获得微软百万名人识别竞赛MS-Cele-1M人脸挑战赛第一名。
6.2010-2012年,与新加坡国立大学颜水成教授联携,三年连续赢得了PASCAL VOC竞赛大奖(ImageNet的前身)。
学术著作
·Scale-Aware Fast R-CNN for Pedestrian Detection.IEEE Trans. Multimedia 20(4): 985-996 (2018)
·Dual-Mode Vehicle Motion Pattern Learning for High Performance Road Traffic Anomaly Detection. CVPR Workshops 2018: 145-152
·Towards Pose Invariant Face Recognition in the Wild. CVPR 2018: 2207-2216
·ThinNet: An Efficient Convolutional Neural Network for Object Detection. ICPR 2018: 836-841
·3D-Aided Deep Pose-Invariant Face Recognition. IJCAI 2018: 1184-1190
·Person re-identification with fusion of hand-crafted and deep pose-based body region features. CoRR abs/1803.10630 (2018)
·Look Across Elapse: Disentangled Representation Learning and Photorealistic Cross-Age Face Synthesis for Age-Invariant Face Recognition. CoRR abs/1809.00338 (2018)
·Detecting The Objects on The Road Using Modular Lightweight Network. CoRR abs/1811.06641 (2018)
·Intention-Net: Integrating Planning and Deep Learning for Goal-Directed Autonomous Navigation. CoRL 2017: 185-194
·High Performance Large Scale Face Recognition with Multi-cognition Softmax and Feature Retrieval. ICCV Workshops 2017: 1898-1906
·Know You at One Glance: A Compact Vector Representation for Low-Shot Learning. ICCV Workshops 2017: 1924-1932
·Audio-visual emotion recognition using deep transfer learning and multiple temporal models. ICMI 2017: 577-582
·Dual-Agent GANs for Photorealistic and Identity Preserving Profile Face Synthesis. NIPS 2017: 65-75
·Driver Head Analysis Based on Deeply Supervised Transfer Metric Learning with Virtual Data. PCM (2) 2017: 286-295
·A Good Practice Towards Top Performance of Face Recognition: Transferred Deep Feature Fusion. CoRR abs/1704.00438 (2017)
·Intention-Net: Integrating Planning and Deep Learning for Goal-Directed Autonomous Navigation. CoRR abs/1710.05627 (2017)
·Optimal depth recovery using image guided TGV with depth confidence for high-quality view synthesis. J. Visual Communication and Image Representation 39: 24-39 (2016)
·Robust Face Recognition with Deep Multi-View Representation Learning. ACM Multimedia 2016: 1068-1072
·Scale-aware Fast R-CNN for Pedestrian Detection. CoRR abs/1510.08160 (2015)
·Combining high resolution color and depth images for dense 3D reconstruction. ICCE-Berlin 2014: 331-334
·Novel edge preserve and depth image recovery method in RGB-D camera systems. ICCE-Berlin 2014: 346-349
·Directional adaptive hole filling for new view synthesis. ICCE 2012: 610-611
·CFA-based motion blur removal using long/short exposure pairs. IEEE Trans. Consumer Electronics 56(2): 332-338 (2010)