[1] 国家自然科学基金面上项目，自主智能系统感知与理解的软硬件协同计算优化与快速验证，2026.1-2029.12
[2] 国家自然科学基金面上项目，无人系统自主定位及环境建模的计算优化与软硬件协同设计，2021.1-2024.12
[3] 国家重点研发计划课题，机器人操作系统总体设计及运行环境（子课题负责人），2017.12-2020.11
[4] 国家重点研发计划课题，场景视觉大数据的透彻感知关键技术（课题骨干），2016.7-2020.6
[5] 广东省科技计划项目，XXXX多媒体处理芯片（联合单位负责人），2017.1-2019.12
[6] 国家自然科学基金面上项目，多粒度可重构视觉处理芯片的设计研究与架构优化，2014.1-2017.12
[7] “863”重点项目，面向领域的并行应用程序综合优化技术与系统，2012.1-2015.12
[8] 部委项目子课题，×××图像压缩算法及其专用ASIC/SoC，2011.1-2015.12
[9] 教育部新世纪优秀人才支持计划，异构多核SoC软硬件协同设计关键技术研究与系统集成，2012.1-2014.12
[10]核高基重大专项分课题，数字电视SoC芯片支撑软件、多媒体应用与多核虚拟化, 2009.1-2011.12
[11]国家自然基金青年基金，多层次并行的可重构视觉处理SoC关键技术研究，2010.1-2012.12
[12]合作课题，专用网络模型、大模型的端侧部署与优化，2025.1-2026.12
[13]合作课题，大规模AI与HPC应用的软硬件协同设计优化，2023.4-2024.4
[14]合作课题，深度网络模型的嵌入式平台计算调度与优化，2020.4-2020.12
[15]合作课题，某嵌入系统实时内核与硬件系统构建，2017.6-2019.12
[16]合作课题，可重构智能处理系统，2017.6-2018.6
[17]合作课题，某处理器汇编与高级语言设计与验证，2015.11-2018.12
[18]合作课题，机顶盒手势控制系统开发，2014.4-2014.12
[19]合作课题，智能视频监控算法在嵌入式平台的设计开发，2012.6-2012.12
[20]合作课题，图像处理与分析库的嵌入式平台移植与优化，2011.11-2012.3
[21]合作课题，嵌入式高清视频解码软件平台，2011.9-2012.1
课题组近期录用与发表的论文：
[1] Huipeng Zhang et al.: DAG-SP: Dual-Attention Guided Semantic Propagation for UAV Video Segmentation with High Temporal Consistency,IEEE Transactions on Circuits and Systems for Video Technology,2026.
[2]Xuhui Wang et al.:Pipelined Microarchitecture for VTA With Circuit and Instruction Co-Optimization，IEEE Transactions on Very Large Scale Integration (VLSI) Systems，2026.
[3] 靳苗苗等，采用置信度分层协同决策的目标检测方法，西安交通大学学报（自然科学版），2026。
[4] Xuyi Yu et al.: Multi-perspective Quantization-Aware Distillation for Semantic Segmentation, IEEE Transactions on Circuits and Systems for Video Technology,2026.
[5] Xuyi Yu et al.: Cross-Head Relation Distillation for Edge-Friendly Semantic Segmentation in Transportation Systems, IEEE Transactions on Consumer Electronics,2026.
[6] Xuyi Yu et al.: Towards Compact Representations: A Region-based Logit Distillation Framework for Edge-Friendly Semantic Segmentation, IEEE Internet of Things Journal,2026.
[7] Huipeng Zhang et al.: DKM-Net: Self-Supervised 3D Occupancy Prediction via Multi-Scale Dynamic Kernel method, Machine Vision and Applications,2026.
[8] Xuyi Yu et al.: Efficient semantic segmentation via logit-guided feature distillation. Neural Networks 199: 108663 (2026).
[9] Ruohan Cheng，et al.: Design of data access and schedule optimization for VTA compiled instruction streams. Future Gener. Comput. Syst. 176: 108165 (2026).
[10] Dong Gou et al.:A Fusion of ReLU and upSample Function With Store in VTA for Higher Throughput of Network Inferencing, Electronics Letters, 2025; 61:e70284.
[11] Xuhui Wang et al.:A bandwidth enhancement method of VTA based on paralleled memory access design. Integr. 94: 102102 (2024).
[12] Feifei Wei et al.:Towards self-explainable graph convolutional neural network with frequency adaptive inception. Pattern Recognit. 146: 109991 (2024).
英文论文更新详见：https://dblp.uni-trier.de/pid/97/597.html
[1]Lu Xu, Chen Hu, Kuizhi Mei: Semi-supervised Regression with Manifold:A Bayesian Deep Kernel Learning Approach. Neurocomputing 497: 76-85 (2022)
[2]Lu Xu, Chen Hu, Ji’an Tao, Jianru Xue, Kuizhi Mei：Improve Regression Network on Depth Hand Pose Estimation with Auxiliary Variable，IEEE Trans. Circuits Syst. Video Techn. 31(3): 890-904 (2021)
[3] Ji Zhang, Kuizhi Mei, Yu Zheng, Jianping Fan：Integrating Part of Speech Guidance for Image Captioning, IEEE Trans. Multim. 23:92-104(2021).
[4] Ji Zhang, Kuizhi Mei, Yu Zheng, Jianping Fan:Learning multi-layer coarse-to-fine representations for large-scale image classification. Pattern Recognition 91: 175-189 (2019)
[5] Ji Zhang, Kuizhi Mei, Yu Zheng, Jianping Fan: Exploiting Mid-Level Semantics for Large-Scale Complex Video Classification. IEEE Trans. Multimedia 21(10): 2518-2530 (2019)
[6] Hao Lei, Kuizhi Mei, Jingmin Xin, Peixiang Dong, Jianping Fan, Hierarchical Learning of Large-Margin Metrics for Large-Scale Image Classification, Neurocomputing 208:46-58 (2016).
[7] Kuizhi Mei, Jinye Peng, Ling Gao, Naiquan (Nigel) Zheng, Jianping Fan:Hierarchical Classification of Large-Scale Patient Records for Automatic Treatment Stratification. IEEE J. Biomedical and Health Informatics 19(4): 1234-1245 (2015)
[8] Jianping Fan, Ji Zhang, Kuizhi Mei, Jinye Peng, Ling Gao: Cost-sensitive learning of hierarchical tree classifiers for large-scale image classification and novel category detection. Pattern Recognition 48(5): 1673-1687 (2015)
[9] Kuizhi Mei, Lu Xu, Boliang Li, Bin Lin, Fang Wang:A real-time hand detection system based on multi-feature. Neurocomputing 158: 184-193 (2015)
[10] Kuizhi Mei, Ji Zhang, Guohui Li, Bao Xi, Nanning Zheng, Jianping Fan: Training more discriminative multi-class classifiers for hand detection. Pattern Recognition 48(3): 785-797 (2015)
[11] Kuizhi Mei,Peixiang Dong,Hao Lei,Jianping Fan, A Distributed Approach for Large-Scale Classifier Training and Image Classification, Neurocomputing 144: 304-317 (2014).
[12] Hao Lei, Kuizhi Mei, Nanning Zheng, Peixiang Dong, Ning Zhou, Jianping Fan: Learning group-based dictionaries for discriminative image representation. Pattern Recognition 47(2): 899-913 (2014) SCI
[13] Peixiang Dong, Kuizhi Mei, Nanning Zheng, Hao Lei, Jianping Fan: Training inter-related classifiers for automatic image classification and annotation. Pattern Recognition 46(5): 1382-1395 (2013) SCI
[14] Kuizhi Mei, Sheng Feng, Guohui Li, Wenlong Chou, Lei Zhang: A virtual touch event method using scene recognition for digital television. IEEE Trans. Consumer Electronics 59(1): 214-219 (2013) SCI
VLSI and multi-processor SoC:
[1] Bin Zhang, Chen Zhao, Kuizhi Mei, Jizhong Zhao, Nanning Zheng:Hierarchical and Parallel Pipelined Heterogeneous SoC for Embedded Vision Processing. IEEE Trans. Circuits Syst. Video Techn.28(6): 1434-1444 (2018)
[2] Yuhai Li, Kuizhi Mei, Yuehu Liu：Improving the Area Efficiency of ACO-Based Routing by Directional Pheromone in Large-Scale NoCs, Microprocessors and Microsystems - Embedded Hardware Design 45: 81-94 (2016).
[3] Bin Zhang, Kuizhi Mei, Jizhong Zhao:Matrix computing coprocessor for an embedded system. Microprocessors and Microsystems - Embedded Hardware Design 39(7): 444-456 (2015)
[4] Bin Zhang, Kuizhi Mei, Nanning Zheng:Coarse-grained Dynamically Reconfigurable Processor for Vision Pre-Processing. Signal Processing Systems 79(1): 45-61 (2015)
[5] Yuhai Li,Kuizhi Mei, Yuehu Liu, Nanning Zheng, Yi Xu:Application-driven dynamic bandwidth allocation for two-layer network-on-chip design,Computers and Electrical Engineering, 40(8):317-332(2014) SCI
[6] Yuhai Li, Kuizhi Mei, Yuehu Liu,Nanning Zheng,Yi Xu, LDBR: Low-Deflection Bufferless Router for Cost-Sensitive Network-on-Chip Design, Microprocessors and Microsystems - Embedded Hardware Design 38(7): 669-680 (2014). SCI
[7] Chen Zhao, Kuizhi Mei, Nanning Zheng: Design of write merging and read prefetching buffer in DRAM controller for embedded processor. Microprocessors and Microsystems - Embedded Hardware Design 38(5): 451-457 (2014) SCI
[8] Bin Zhang, Kuizhi Mei, Nanning Zheng: Reconfigurable Processor for Binary Image Processing. IEEE Trans. Circuits Syst. Video Techn. 23(5): 823-831 (2013) SCI
[9] Yuhai Li, Kuizhi Mei and Peixiang Dong, “An Efficient and Low Memory Requirement Algorithm for Extracting Image Component Information”, pp.255-267, Vol.3 No. 2, July 2011, International Journal of Advanced Intelligence.
[10] Kuizhi Mei, Bin Zhang and Chenyang Ge, A Hierarchical and Parallel SoC Architecture for Vision Processor, IEICE Electronics Express, Vol. 6, No.19, pp:1380-1386，Oct. 10 2009. SCI
[11] K. Mei, N. Zheng, C. Huang, Y. Liu, and Q. Zeng, “VLSI Design of A High-Speed and Area-Efficient JPEG2000 Encoder,” IEEE Trans. Circuits Syst. Video Technol., vol. 17, no. 8, pp:1065-1078, Aug. 2007. SCI
[12] K. Mei, N. Zheng and Huub van de Wetering, High-speed and Memory-efficient VLSI design of 2D DWT for JPEG2000, Electronics Letters, Vol. 42 No.16. Pages:907-908. Aug. 2006. SCI
[1] Pengfei Qu, Shenliang Li, Jinyang Zhang, Zhansheng Duan and Kuizhi Mei，A Low-cost and Robust Mapping and Relocalization Method Base on Lidar Inertial Odometry，2021 IEEE International Conference on Real-time Computing and Robotics(RCAR 2021)
[2] Jinyang Zhang, Pengfei Qu, Shenliang Li, Kuizhi Mei and Zhansheng Duan，Colorful Reconstruction from Solid-State-LiDAR and Monocular Version，2021 IEEE International Conference on Real-time Computing and Robotics (RCAR 2021)
[3] Jianyu Yan,Kuizhi Mei,Attention Residual Network with 3D convolutional neural network for 3D Human Pose Estimation., 2021 IEEE International Conference on Real-time Computing and Robotics(RCAR 2021)
[4] Ji'an Tao, Lu Xu, Xinyan Ma, Jianyu Yan and Kuizhi Mei:Scene-Perception Graph Convolutional Networks for Human Action Prediction, International Joint Conference on Neural Networks (IJCNN 2021)
[5] Fan Chang, Jun Cheng, Jianling Yang, Hanting Huang, Kuizhi Mei, Multi-channel signal processing heterogeneous microsystem based on FPGA and Application Processor，2020 IEEE 15th International Conference on Solid-State and Integrated Circuit Technology (ICSICT).
[6]Hejie Yu, Jun Cheng, Xiangnan Zhang, Yuzhe Gao, Kuizhi Mei, Implementation of Convolutional Neural Network with Co-design of High-Level Synthesis and Verilog HDL, 2020 IEEE 15th International Conference on Solid-State and Integrated Circuit Technology (ICSICT).
[7]Ji'An Tao, Lu Xu, Xinyan Ma and Kuizhi Mei: An Efficient System for Predicting Hand-Object Contact Probability Based on RGB Image Sequences, 12th International Conference on Advanced Computational Intelligence (ICACI2020).
[8]Yinqi Li, Lu Xu, Weihua Shu, Ji'An Tao, Kuizhi Mei: AutoGesNet: Auto Gesture Recognition Network Based on Neural Architecture Search, 12th International Conference on Advanced Computational Intelligence (ICACI2020).
[9]Jianyu Yan，Lu Xu，Kuizhi，Mei：Optical Flow Prediction in Auto Driving from Single Image via Conditional Variational AutoEncoder, 2019 IEEE International Conference on Unmanned Systems and Artificial Intelligence (ICUSAI).
[11]Yafei Li, Kuizhi Mei, Xiao Wang, Zeng Zhang, Hejie Yu： Collaborative Implementation of Hardware-Oriented GBDT Compress Algorithm Based on DSP+FPGA, 2019 IEEE International Conference on ASIC (ASICON) 2019: 1-4.
[12]Xiao Wang, Yafei Li, Yinqi Li, Kuizhi Mei: A Novel Compression Algorithm for Hardware-Oriented Gradient Boosting Decision Tree Classification Model. International Conference on Intelligent Computing (ICIC) (3) 2019: 381-391.
[13]Lu Xu, Chen Hu, Yinqi Li, Ji'an Tao, Jianru Xue, Kuizhi Mei: Deep Conditional Variational Estimation for Depth-Based Hand Poses. Automatic Face and Gesture Recognition (FG) 2019: 1-7.
[14]Ji Zhang, Kuizhi Mei, Xiao Wang, Yu Zheng Jianping Fan: From Text to Video: Exploiting Mid-Level Semantics for Large-Scale Video Classification, International Conference on Pattern Recognition (ICPR) 2018.
[15]Haibo Zhao, Fei Zheng, Jian Wu, Baosong Nan, Boliang Li, Kuizhi Mei: Automatic Parallelization for Binary on Multi-core Platforms，The 2nd International Conference on Computer Science and Application Engineering 2018.
[16]Xu L, Zhang L, Wang K, et al. An efficient fast hand tracking approach based on segmentation[C]. Chinese Association of Automation (YAC), Youth Academic Annual Conference of. IEEE, 2016: 19-24.
[17]Ji Zhang, Kuizhi Mei*, Jianping Fan, Learning Multi-Layer Coarse-to-Fine Representations for Over 10,000 Image Categories, 2016, KDD workshop on Large-Scale Deep Learning on Data Mining.
[18]Peixiang Dong, Kuizhi Mei, Ji Zhang, Hao Lei, Jianping Fan: Task-Driven Image Retrieval Using Geographic Information. MMM (2) 2014: 214-226
[19]Yonglong Zhou, Kuizhi Mei, Xiang Ji, Peixiang Dong: Parallelization and Optimization of SIFT on GPU Using CUDA. HPCC/EUC 2013: 1351-1358
[20]Lei Xie, Kuizhi Mei, Yuhai Li: REPAIR: A Reliable Partial-Redundancy-Based Router in NoC. NAS 2013: 173-177
[21]Hao Lei, Ning Zhou, Kuizhi Mei, Peixiang Dong, Jianping Fan: Constructing Hierarchical Visual Tree for Discriminative Image Representation and Classification. PCM 2013: 637-648
[22]Li Yuhai, Xie Lei, Mei Kuizhi and Liu Yuehu, “R+S: an application-specific hybrid on-chip network based on self-adaptive optimization for communication”, 2012 IEEE International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS 2012).
部分中文期刊：
[1] 张良，张增，舒伟华，梅魁志，基于YOLOv3的卷积层结构化剪枝，计算机工程与应用, Vol. 57 (6): 131-137,2021
[2] 高榕，张良，梅魁志，基于Caffe的嵌入式多核处理器深度学习框架并行实现，西安交通大学学报，Vol 52 No.6(2018.06),36-41.
[3] 王锟, 宋永红, 郑斐, 梅魁志. 基于卷积神经网络的嵌入式手势检测算法. 计算机工程与应用. 2019 Vol. 55 (4): 137-141.
[4] 仲济源，梅魁志，温哲西, GIST特征提取的异构并发流计算实现，计算机工程与应用，2015，51（6）：139-144.
[5] 崔继岳，梅魁志，刘冬冬，李博良，面向OpenCL的Mali GPU仿真器构建研究，西安交通大学学报，Vol 49 No.2(2015.02),20-24.
[6] 丑文龙,梅魁志,高增辉,李博良，ARM GPU的多任务调度设计与实现,西安交通大学学报，Vol.48 No.12 (2014.12)，87-92.
部分授权国家发明专利
[1] 一种基于软硬件结合方式的大容量快速 NandFlash 存储实现方法，申请号：202010044024.2
[2] 基于点云配准ICP算法构建K-D树的方法，申请号：201910350982.X.
[3] 基于FPGA和行输出优先的卷积神经网络实现方法，ZL201910323955.3.
[4] 卷积神经网络全连接层稀疏的权值矩阵的压缩方法及系统，ZL201910339608X．
[5] 一种适用于专用处理器实时控制的软件编程模式及其汇编链接方法，ZL201810569188.X.
[6] 一种针对多核平台的二进制程序自动并行化的设计方法，ZL201510657166.5
[7] ARM 二进制代码的NEON 向量化转换设计，ZL201510574950.X
[8] 基于共享弱分类器组合的人手分类器以及训练和检测方法，ZL201410351088.1
[9] 基于场景识别的触摸屏技术专利，ZL201210065207.8
[10] 支持linux的硬件抽象层的虚拟网卡设计方法，ZL201110175932.6
[11] ARM虚拟机中基于MMU的外设访问控制的实现方法，ZL201110217684.7）
[12] 一种面向嵌入式hypervisor的中断虚拟化操作方法，ZL201110191645.3
[13] 一种面向嵌入式Hypervisor的多视窗显示虚拟化操作方法，ZL201110191642.0
[14] 复用客户操作系统设备驱动的实现方法，ZL 201110255829.2
[15] 可重构二值运算器、可重构二值图像处理系统及其实现基本形态学算法的方法，ZL201110252219.7
[16] 矩阵处理器及其指令集和嵌入式系统，ZL201110303919.4。
[17] 一种可重构的视觉前处理器和视觉处理系统，ZL 201210047720.4
[18] 一种用于视频图像的非均匀采样方法，ZL200910021270.x
[19] 一种双总线的视觉信息处理芯片架构，ZL200910021723.9
[20] 一种面向写穿透Cache的SDRAM读写方法 ，ZL200810232131.7
[21] 流输入读取的多媒体扩展指令的设计与实现方法，ZL200610105066.2
[22] 一种基于寄存器组的行程解码与反扫描实现方法，ZL200610042757.2,
[23] 同步流水算术编码器的VLSI实现方法，ZL200410026019.X
[24] 一种0冗余编码时钟的位平面编码器的VLSI实现方法，ZL200410026018.5
[25] 内在并行的二维离散小波变换的VLSI结构设计方法，ZL03114602.3
PCT申请
[1] 一种集成第三方图形库到操作系统中的方法，申请号PCT/CN2012/072868
[2] Method of Scene Recognition Based Virtual Touch Event，申请号PCT/CN2012/073434
所带领的项目组围绕视频图像与视觉信息处理、深度网络模型加速、多核SoC芯片设计与嵌入式系统软件应用已经完成的芯片与软件系统：
1) 已构建从网络模型编译到高精度FPGA部署的全流程软硬件系统实现（TVM+张量处理器PVTA或专用卷积加速器），推理性能>NVDLA，在TVM开源社区发布多个优化与改进设计。
2) 具有自主产权的符合JPEG2000标准的编解码IP核设计与验证以及基于小波的新一代可伸缩视频解码FPGA系统验证。
3）异构多核媒体处理器DTV200A：支持MPEG2标清视频的实时解码与显示以及uCLinux＋MiniGUI的应用程序与交互。
4）基于异构多核架构的视觉处理器系列芯片：单芯片可完成实时目标跟踪、3D位姿计算等机器视觉应用
5）异构多核SoC芯片的系统支撑软件-嵌入式虚拟机Xtimes，将操作系统作为超级进程，可直接控制资源，支持强实时环境构建。
6）二进制应用的自动并行化软件工具链：在对串行二进制程序的热点函数翻译为LLVM IR后进行ARM多核、SIMD(NEON)等并行化处理。
7）面向大数据应用的新型学习算法与并行/分布式计算实现研究
