劉宏

劉宏，工學博士、北京大學教授、博士生導師。國家級領軍人才，科技部國家重點研發計劃“智能機器人”總體組專家，中國人工智能學會副理事長，深圳市第六屆政協委員 ^[1]  。

長期從事計算機視覺與智能機器人、機器學習與智能人機交互等領域的教學和科研工作。曾師從蔡鶴皋院士、王選院士、石青雲院士等知名專家。多次赴美國、加拿大、日本和新加坡等國家的多所著名大學和研究機構訪問交流。2016年—2020年擔任Elsevier、IET相繼出版的SCI檢索國際學術期刊CAAI Transactions on Intelligence Technology創刊主編，現任執行主編。 ^[4] 

先後承擔 20餘項國家級重要科研項目，包括863課題、 973課題、國家重點研發計劃項目、新一代人工智能重大項目和國家自然科學基金重點項目等。已發表學術論文200多篇，近年的研究成果被國內外同行研究機構引用7000餘次，相關成果申報/獲得國家發明專利40餘項 ^[6]  。獲國家航天科技進步獎、吳文俊人工智能科學技術獎、日內瓦國際發明博覽會獎、北京大學教學優秀獎、安泰獎和北京大學“十佳教師”候選人等榮譽 ^[1]  。

作為深圳教育界首位國家級領軍人才，二十多年來劉宏教授專注於所熱愛的人工智能和機器人事業，先後培養了100多名碩士博士研究生和博士後 ^[1]  ，是我國“智能科學與技術”新學科建設的積極倡導者、智能機器人領域科技創新的潛心實踐者。“十四五”期間劉宏教授的研究特色將聚焦在“機器人視覺感知與自主學習”等領域 ^[1]  。

近年主持/承擔的主要科研項目 ^[1]  :

·國家新一代人工智能重大項目：基於數字孿生的室內服務機器人自主學習與進化關鍵技術（在研）

·國家重點研發計劃項目：面向智能手機制造的柔性機器人自動化生產線研製及示範應用（在研、 指導）

·國家自然科學基金項目：面向混合增強智能運動規劃的機器人位姿空間建模方法（在研）

·深圳市基礎研究重點項目：基於情感計算人和機器人交互模型研究 （在研）

·深圳市高等院校穩定支持計劃重點項目：面向複雜場景機器人高效作業的混合增強智能（在研）

·國家自然科學基金(聯合基金)重點項目：面向服務機器人的視聽感知融合與多模態人機交互關鍵技術 （已完成）

·國家自然科學基金項目：基於麥克風陣列的移動機器人實時聲源定位方法研究（已完成）

·國家自然科學基金項目：人機互動環境下機器人實時運動規劃研究（已完成）

·國家自然科學基金項目：面向人體目標實時跟蹤的視覺注意轉移機制研究（已完成）

·國家863課題：面向HRI的機器人視聽覺注意機制及運動規劃技術（已完成）

·國家863課題：基於多源信息融合的交通事件自動檢測技術（已完成）

·科技部創新人才推進計劃：面向複雜場景人機交互的嵌入式仿生視覺技術（已完成）

·教育部博士點基金課題：面向顯著事件主動感知的仿生立體視覺研究（已完成）

·廣東省重大產業攻關項目：新一代家用服務機器人關鍵技術突破及集成應用示範（已完成）

·深圳市“創新鏈-產業鏈”雙鏈融合重大產業化項目：智能電視生產流水線的視覺檢測和定位方法 （已完成）

·深圳市戰略新興產業項目：網絡環境下智能監控系統公共技術平台（已完成）

·深圳市戰略新興產業項目：物聯網智能感知技術工程實驗室（已完成）

·深圳市基礎研究重點項目：面向複雜場景人機交互的仿生視覺技術與系統（已完成）

近年發表的代表性論文: ^[1] 

[1] Wenhao Li, Hong Liu(劉宏), Runwei Ding,Mengyuan Liu, Pichao Wang and Wenming Yang. Exploiting temporal contexts withstrided transformer for 3d human pose estimation. IEEE Transactions onMultimedia(TMM). Accepted in 2022.（多媒體領域頂級國際學術期刊）

[2] Wei Shi, Hong Liu(劉宏), and MengyuanLiu. Image-to-video person re-identification using three-dimensional semanticappearance alignment and cross-modal interactive learning. Pattern Recognition(PR),2022, 122: 108314.（模式識別領域頂級國際學術期刊）

[3] Weibo Huang and Hong Liu(劉宏). A robustpixel-aware gyro-aided KLT feature tracker for large camera motions. IEEE Transactionson Instrumentation and Measurement(TIM), 2022, 71: 1-14. (機器智能領域重要國際學術期刊)

[4] Bing Yang, Hong Liu(劉宏),and Xiaofei Li. Learning deep direct-path relative transfer function forbinaural sound source localization. IEEE/ACM Transactions on Audio, Speech, andLanguage Processing(TASLP), 2021, 29: 3491-3503.（機器聽覺領域頂級國際學術期刊）

[5] Meijia Song, Hong Liu(劉宏), Wei Shi, and XiaLi. PCLoss: Fashion Landmark Estimation with Position Constraint Loss. PatternRecognition(PR), 2021, 118: 108028.（模式識別領域頂級國際學術期刊）

[6] Hao Tang, Hong Liu(劉宏), Dan Xu, PhilipHS Torr, and Nicu Sebe. Attentiongan: Unpaired image-to-image translation usingattention-guided generative adversarial networks. IEEE Transactions on NeuralNetworks and Learning Systems(TNNLS), 2021, 1-16.（人工智能領域頂級國際學術期刊）

[7] Hanrong Ye, Hong Liu(劉宏), Fanyang Meng,and Xia Li. Bi-Directional Exponential Angular Triplet Loss for RGB-InfraredPerson Re-Identification. IEEE Transactions on Image Processing(TIP),2021, 30:1583-1595. （人工智能領域頂級國際學術期刊）

[8] Hao Tang, Hong Liu(劉宏), Wei Xiao, andNicu Sebe. When Dictionary Learning Meets Deep Learning: Deep DictionaryLearning and Coding Network for Image Recognition With Limited Data. IEEETransactions on Neural Networks and Learning Systems(TNNLS),2021, 32(5): 2129-2141.（人工智能領域頂級國際學術期刊）

[9] Weibo Huang, Hong Liu(劉宏), and Weiwei Wan.An Online Initialization and Self-Calibration Method for Stereo Visual-InertialOdometry. IEEE Transactions on Robotics(TRO), 2020, 36(4):1153-1170.（機器人領域頂級國際學術期刊）

[10] Jie Wen, Yong Xu, and Hong Liu(劉宏). IncompleteMultiview Spectral Clustering With Adaptive Graph Learning. IEEE Transactionson Cybernetics(TCYB), 2020, 50(4): 1418-1429.（人工智能領域頂級國際學術期刊）

[11] Hao Tang, Hong Liu(劉宏), and Nicu Sebe. Unified Generative Adversarial Networks for ControllableImage-to-Image Translation. IEEE Transactions on Image Processing(TIP),2020, 29: 8916-8929.（人工智能領域頂級國際學術期刊）

[12] Jiayao Ma, Weiwei Wan, Kensuke Harada,Qiuguo Zhu, and Hong Liu(劉宏). Regrasp Planning Using Stable ObjectPoses Supported by Complex Structures. IEEE Transactions on Cognitive andDevelopmental Systems(TCDS), 2019, 11(2): 257-269.（人工智能領域重要國際學術期刊）

[13] Bing Yang, Hong Liu(劉宏), Cheng Pang andXiaofei Li. Multiple Sound Source Counting and Localization Based on TF-WiseSpatial Spectrum Clustering. IEEE/ACM Transactions on Audio, Speech andLanguage Processing(TASLP), 2019, 27(8):1241-1255.（機器聽覺領域頂級國際學術期刊）

[14] Fanyang Meng, Hong Liu(劉宏), YongshengLiang, Juanhui Tu and Mengyuan Liu. Sample Fusion Network: An End-to-End DataAugmentation Network for Skeleton-based Human Action Recognition. IEEETransactions on Image Processing(TIP), 2019, 28(11): 5281-5295.（人工智能領域頂級國際學術期刊）

[15] Mengyuan Liu, Hong Liu(劉宏), and Chen Chen, 3D action recognition using multi-scale energy-based global ternary image, Accepted by IEEE Transactions on Circuits and Systems for Video Technology(TCSVT), 28(8): 1824-1838 (2018) （多媒體領域頂級國際學術期刊）

[16] Mengyuan Liu, Hong Liu(劉宏), Chen Chen, Robust 3D Action Recognition Through Sampling Local Appearances and Global Distributions. IEEE Transaction on Multimedia(TMM) 20(8): 1932-1947 (2018) （多媒體領域頂級國際學術期刊）

[17] Chen Pang, Hong Liu(劉宏), Jie Zhang, Binaural Sound Localization Based on Reverberation Weighting and Generalized Parametric Mapping, Accepted by IEEE Transaction on Audio, Speech and Language Processing(TASLP) VOL.25:1618-1632, 2017（機器聽覺領域頂級國際學術期刊）

[18] Pingping Wu, Hong Liu, Chao Xu, Yuan Gao, Zheyuan Li, Xuewu Zhang, How do you smile? Towards a comprehensive smile analysis system. Neurocomputing 235: 245-254 (2017) （人工智能領域重要國際學術期刊）

[19] Mengyuan Liu, Hong Liu(劉宏), and Chen Chen, Enhanced skeleton visualization for view invariant human action recognition, Pattern Recognition(PR) 68: 346-362 (2017)（模式識別領域頂級國際學術期刊）

[20] Qianru Su, Hong Liu(劉宏), Tatsuya Harada, Online Growing Neural Gas for Anomaly Detection in Changing Surveillance Scenes, Pattern Recognition(PR) 64: 187-201, 2017.（模式識別領域頂級國際學術期刊）

[21] Pingping Wu, Hong Liu (劉宏), Xiaofei Li, A Novel Lip Descriptor for Audio-Visual Keyword Spotting Based on Adaptive Decision Fusion, IEEE Transactions on Multimedia (TMM). vol.18(3): 326-338, 2016. （多媒體領域頂級國際學術期刊）

[22] Qianru Sun, Hong Liu, Liqian Ma, Tianwei Zhang, A novel hierarchical Bag-of-Words model for compact action representation. Neurocomputing 174: 722-732 (2016) （人工智能領域重要國際學術期刊）

[23] Yuan Gao, Hong Liu(劉宏), Pingping Wu P. A new descriptor of gradients Self-Similarity for smile detection in unconstrained scenarios. Neurocomputing. vol.174: 1077-1086, 2016. （人工智能領域重要國際學術期刊）

[24] Mengyuan Liu, Hong Liu(劉宏), Depth Context: a new descriptor for human activity recognition by using sole depth sequences. Neurocomputing. vol.175: 747-758, 2016. （人工智能領域重要國際學術期刊）

[25] Qianru Sun Q, Hong Liu(劉宏), Liqian Ma, A novel hierarchical Bag-of-Words model for compact action representation. Neurocomputing. vol.174: 722-732, 2016. （人工智能領域重要國際學術期刊）

[26] Jie Zhang, Hong Liu (劉宏), Robust Acoustic Localization Via Time-Delay Compensation and Interaural Matching Filter, IEEE Transactions on Signal Processing(TSP), vol.63, no.18, pp. 4771-4783, 2015. （多媒體信號處理領域頂級國際學術期刊）

[27] Qianru Sun, Hong Liu(劉宏), Mengyuan Liu, Human activity prediction by mapping grouplets to recurrent Self-Organizing Map. Neurocomputing, vol.177: 427-440, 2015. （人工智能領域重要國際學術期刊）

[28] Wenhao Li, Hong Liu, Hao Tang, Pichao Wang, and Luc VanGool. MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimation.Accepted by IEEE Conference on Computer Vision and Pattern Recognition (CVPR),2022. (計算機視覺領域頂級國際學術會議)

[29] Tianyu Guo, Hong Liu(劉宏), Zhan Chen, Mengyuan Liu, Tao Wang,Runwei Ding. Contrastive Learning from Extremely Augmented Skeleton Sequencesfor Self-supervised Action Recognition. Accepted by Proceedings of the AAAIConference on Artificial Intelligence (AAAI), 2022. （人工智能領域頂級國際學術會議）

[30] Tao Wang, Hong Liu(劉宏), Pinhao Song, Tianyu Guo, Wei Shi.Pose-guided Feature Disentangling for Occluded Person Re-identification basedon Transformer. Accepted by Proceedings of the AAAI Conference on ArtificialIntelligence (AAAI), 2022.（人工智能領域頂級國際學術會議）

[31] Yidi Li, Hong Liu(劉宏), Hao Tang. Multi-Modal PerceptionAttention Network with Self-Supervised Learning for Audio-Visual SpeakerTracking. Accepted by Proceedings of the AAAI Conference on ArtificialIntelligence (AAAI), 2022.（人工智能領域頂級國際學術會議）

[32] Zhan Chen, Sicheng Li, Bing Yang, Qinghan Li, and HongLiu(劉宏).Multi-Scale Spatial Temporal Graph Convolutional Network for Skeleton-BasedAction Recognition. Proceedings of the AAAI Conference on ArtificialIntelligence (AAAI), 2021, 35(2): 1113-1122.（人工智能領域頂級國際學術會議）

[33] Wanlu Xu, Hong Liu(劉宏), Wei Shi, Ziling Miao, Zhisheng Lu, andFeihu Chen, Adversarial Feature Disentanglement for Long-Term PersonRe-identification, International Joint Conference on Artificial Intelligence (IJCAI),2021.（人工智能領域頂級國際學術會議）

[34] Ziling Miao, Hong Liu(劉宏), Wei Shi, Wanlu Xu, and Hanrong Ye,Modality-aware Style Adaptation for RGB-Infrared Person Re-Identification,International Joint Conference on Artificial Intelligence (IJCAI),2021: 19-27.（人工智能領域頂級國際學術會議）

[35] Peng Wei, Guoliang Hua, Weibo Huang, Hong Liu(劉宏). Unsupervised MonocularVisual-inertial Odometry Network, International Joint Conferences on ArtificialIntelligence Organization (IJCAI), 2020: 2347-2354.（人工智能領域頂級國際學術會議）

[36] Jie Wen, Zheng Zhang, Yong Xu, Bob Zhang, Lunke Fei, and HongLiu(劉宏).Unified Embedding Alignment with Missing Views Inferring for IncompleteMulti-View Clustering. Proceedings of the AAAI Conference on ArtificialIntelligence (AAAI), 2019, 33(01): 5393-5400.（人工智能領域頂級國際學術會議）

[37] Xia Li, Zhisheng Zhong, Jianlong Wu, Yibo Yang, ZhouchenLin, Hong Liu(劉宏). Expectation-maximization attention networks for semanticsegmentation, International Conference in Computer Vision (ICCV),2019: 9167-9176.(計算機視覺領域頂級國際學術會議)

[38] Xia Li, Jianlong Wu, Zhouchen Lin, Hong Liu(劉宏), Hongbin Zha. Recurrent squeeze-and-excitationcontext aggregation net for single image deraining, European Conference onComputer Vision (ECCV), 2018: 254-269.(計算機視覺領域頂級國際學術會議)

[39] Weibo Huang, Hong Liu(劉宏). Online initialization and automatic camera-IMUextrinsic calibration for monocular visual-inertial SLAM, InternationalConference on Robotics and Automation (ICRA), 2018: 5182-5189.(智能機器人領域頂級國際學術會議)

[40] Dan Xu, Wei Wang, Hao Tang, Hong Liu(劉宏), Nicu Sebe, Elisa Ricci.Structured attention guided convolutional neural fields for monocular depth estimation,IEEE Conference on Computer Vision and Pattern Recognition (CVPR),2018: 3917-3925.(計算機視覺領域頂級國際學術會議)

[41] Hao Tang, Hong Liu(劉宏), Wei Xiao, A Novel Feature MatchingStrategy for Large Scale Image Retrieval, International Joint Conference on ArtificialIntelligence (IJCAI), 2016: 2053-2059. （人工智能領域頂級國際學術會議）

[42] Chen Chen, Mengyuan Liu, Baochang Zhang, Jungong Han,Junjun Jiang, and Hong Liu(劉宏). 3D Action Recognition Using Multi-temporal DepthMotion Maps and Fisher Vector, International Joint Conference on ArtificialIntelligence (IJCAI), 2016: 3331-3337（人工智能領域頂級國際學術會議）

[43] Hong Liu(劉宏), Mengdi Yue, and Jie Zhang. Probabilistic BinauralMultiple Sources Localization Based on Time-delay Compensation EstimatorClustering Analysis, IEEE International Conference on Intelligent Robots andSystems (IROS), 2016: 4537-4544. （智能機器人領域重要國際學術會議）

[44] Hong Liu(劉宏), Fang Xiao and Can Wang. A predictive model for narrowpassage path planner by using Support Vector Machine in changing environments,IEEE International Conference on Robotics and Automation (ICRA),2015: 2991-2996.(智能機器人領域頂級國際學術會議)

[45] Hong Liu(劉宏), Cheng Pang and Jie Zhang. Binaural sound sourcelocalization based on generalized parametric model and two-layer matchingstrategy in complex environments, IEEE International Conference on Robotics andAutomation (ICRA), 2015: 4496-4503.(智能機器人領域頂級國際學術會議)

作為北京大學首批、深圳教育界首位國家級領軍人才 ^[5]  ，二十多年來劉宏教授專注於所熱愛的人工智能和機器人事業，先後培養了100多名碩士博士研究生和博士後，是我國“智能科學與技術”新學科建設的積極倡導者參與者、智能機器人領域科技創新的潛心實踐者。 ^[1]