摘要： 针对SIFT特征提取算法过程复杂且实时性低的缺陷, 提出了一种基于GPU的实时尺度不变特征变换(Scale-inva-riant feature transform, SIFT)的优化算法——CUDA Optimized SIFT(CoSift)。该算法首先利用CUDA流并发构建SIFT尺度空间, 在此过程中充分利用了CUDA存储器模型中的高速存储器来提高数据访问速度, 并对二维高斯卷积核进行降维优化来减少计算量, 然后设计了基于warp的直方图算法策略, 重新平衡了特征描述过程中的工作负载。其与CPU版本的常用算法及GPU版本的改进算法的对比实验表明, CoSift算法在未降低特征提取准确性的前提下, 极大地提高了实时性能, 且对大尺寸图像有相对更高的优化效果, 在使用单块 GTX 1080Ti的GPU环境下, 该算法可以在7.7~8.6ms(116.28~129.87fps)内提取出关键点。CoSift算法有效降低了传统SIFT算法过程的复杂性, 提升了实时性能, 能较好地应用于对SIFT算法实时性要求较高的场景。
中图分类号:
| [1]LOWE D G.Distinctive Image Features from Scale-Invariant Keypoints[J].International Journal of Computer Vision, 2004, 60(2):91-110. [2]KE N Y, SUKTHANKAR R.PCA-SIFT:A More Distinctive Representation for Local Image Descriptors[C]∥Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.IEEE Computer Society, 2004. [3]BAY H, TUYTELAARS T, GOOL L V.SURF:Speeded UpRobust Features[C]∥European Conference on Computer Vision.2006. [4]DU C, YUAN J, DONG J, et al.GPU based Parallel Optimization for Real Time Panoramic Video Stitching[J].Pattern Recognition Letters, 2018, 133(5):62-69. [5]ACHARYA K A, BABU R V, VADHIYAR S S.A Real-Time Implementation of SIFT Using GPU[J].Journal of Real-Time Image Processing, 2014, 14(8):267-277. [6]ZHOU Y, MEI K, XIANG J, et al.Parallelization and Optimization of SIFT on GPU Using CUDA[C]∥IEEE International Conference on High Performance Computing & Communications & IEEE International Conference on Embedded & Ubiquitous Computing.2014. [7]LI Z, JIA H, ZHANG Y.HartSift:A High-Accuracy and Real-Time SIFT Based on GPU[C]∥2017 IEEE 23rd International Conference on Parallel and Distributed Systems (ICPADS).IEEE Computer Society, 2017. [8]NVIDIA Corporation.CUDA Programming Guide 9.0[OL].https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html. [9]NVIDIA Corporation.CUDA Toolkit Documentation v9.0[OL].https://docs.nvidia.com/cuda/cuda-c-programming-guide. [10]TIAN W, XU F, WANG H Y, et al. Fast Scale Invariant Feature Transform Algorithm Based on CUDA.Computer Engineering, 2010, 36(8):219-221. [11]YAN J H, HANG Y Q, XU J F, et al.Quick Realization of CUDA-Based Registration of High-Resolution Digital Video Images[J].Chinese Journal of Scientific Instrument, 2014, 35(2):380-386. [12]RAGHU R P K, SURESH M, JOHN M.An Approach to Parallelization of SIFT Algorithm on GPUs for Real-Time Applications[J].Journal of Computer and Communications, 2016, 4(17):18-50. [13]JIANG C, GENG Z X, LOU B, et al.Parallel Processing Re-search on SIFT Feature Matching Algorithm Based on GPU[J].Computer Science, 2013, 40(12):295-297, 307. [14]WU C.SiftGPU :A GPU Implementation of Scale InvariantFeature Transform (SIFT)[OL].http://cs.unc.edu/~ccwu/siftgpu. [15]BJRKMAN M, BERGSTRM N, KRAGIC D.Detecting, Segmenting and Tracking Unknown Objects Using Multi-label MRF Inference[J].Computer Vision and Image Understanding, 2014, 118:111-127. [16]ZHANG K, YANG H Y, SHI L Y.Panorama Generation ofSIFT and Stitch Line Based on CUDA[J].Computer Technology and Development, 2015(9):22-26. [17]ZHI X, YAN J, HANG Y, et al.Realization of CUDA-BasedReal-Time Registration and Target Localization for High-Resolution Video Images[J].Journal of Real-Time Image Proces-sing, 2016, 16:1025-1036. [18]The Oxford Buildings Dataset[OL].http://www.robots.ox.ac.uk/~vgg/data/oxbuilding. |
| [1] | 张源, 康乐, 宫朝辉, 张志鸿. 基于Bi-LSTM的期货市场关联交易行为检测方法 Related Transaction Behavior Detection in Futures Market Based on Bi-LSTM 计算机科学, 2022, 49(7): 31-39. https://doi.org/10.11896/jsjkx.210400304 |
| [2] | 曾志贤, 曹建军, 翁年凤, 蒋国权, 徐滨. 基于注意力机制的细粒度语义关联视频-文本跨模态实体分辨 Fine-grained Semantic Association Video-Text Cross-modal Entity Resolution Based on Attention Mechanism 计算机科学, 2022, 49(7): 106-112. https://doi.org/10.11896/jsjkx.210500224 |
| [3] | 程成, 降爱莲. 基于多路径特征提取的实时语义分割方法 Real-time Semantic Segmentation Method Based on Multi-path Feature Extraction 计算机科学, 2022, 49(7): 120-126. https://doi.org/10.11896/jsjkx.210500157 |
| [4] | 刘伟业, 鲁慧民, 李玉鹏, 马宁. 指静脉识别技术研究综述 Survey on Finger Vein Recognition Research 计算机科学, 2022, 49(6A): 1-11. https://doi.org/10.11896/jsjkx.210400056 |
| [5] | 汪晋, 刘江. 基于GPU的并行DILU预处理技术 GPU-based Parallel DILU Preconditioning Technique 计算机科学, 2022, 49(6): 108-118. https://doi.org/10.11896/jsjkx.210300259 |
| [6] | 高元浩, 罗晓清, 张战成. 基于特征分离的红外与可见光图像融合算法 Infrared and Visible Image Fusion Based on Feature Separation 计算机科学, 2022, 49(5): 58-63. https://doi.org/10.11896/jsjkx.210200148 |
| [7] | 徐涛, 陈奕仁, 吕宗磊. 基于改进YOLOv3的机坪工作人员反光背心检测研究 Study on Reflective Vest Detection for Apron Workers Based on Improved YOLOv3 Algorithm 计算机科学, 2022, 49(4): 239-246. https://doi.org/10.11896/jsjkx.210200119 |
| [8] | 李嘉睿, 凌晓波, 李晨曦, 李子木, 杨家海, 张蕾, 吴程楠. 基于贝叶斯攻击图的动态网络安全分析 Dynamic Network Security Analysis Based on Bayesian Attack Graphs 计算机科学, 2022, 49(3): 62-69. https://doi.org/10.11896/jsjkx.210800107 |
| [9] | 左杰格, 柳晓鸣, 蔡兵. 基于图像分块与特征融合的户外图像天气识别 Outdoor Image Weather Recognition Based on Image Blocks and Feature Fusion 计算机科学, 2022, 49(3): 197-203. https://doi.org/10.11896/jsjkx.201200263 |
| [10] | 耿海军, 王威, 尹霞. 基于混合软件定义网络的单节点故障保护方法 Single Node Failure Routing Protection Algorithm Based on Hybrid Software Defined Networks 计算机科学, 2022, 49(2): 329-335. https://doi.org/10.11896/jsjkx.210100051 |
| [11] | 任首朋, 李劲, 王静茹, 岳昆. 基于集成回归决策树的lncRNA-疾病关联预测方法 Ensemble Regression Decision Trees-based lncRNA-disease Association Prediction 计算机科学, 2022, 49(2): 265-271. https://doi.org/10.11896/jsjkx.201100132 |
| [12] | 张师鹏, 李永忠. 基于降噪自编码器和三支决策的入侵检测方法 Intrusion Detection Method Based on Denoising Autoencoder and Three-way Decisions 计算机科学, 2021, 48(9): 345-351. https://doi.org/10.11896/jsjkx.200500059 |
| [13] | 冯霞, 胡志毅, 刘才华. 跨模态检索研究进展综述 Survey of Research Progress on Cross-modal Retrieval 计算机科学, 2021, 48(8): 13-23. https://doi.org/10.11896/jsjkx.200800165 |
| [14] | 张丽倩, 李孟航, 高珊珊, 张彩明. 面向计算机辅助舌诊关键问题的解决方案综述 Summary of Computer-assisted Tongue Diagnosis Solutions for Key Problems 计算机科学, 2021, 48(7): 256-269. https://doi.org/10.11896/jsjkx.200800223 |
| [15] | 暴雨轩, 芦天亮, 杜彦辉, 石达. 基于i_ResNet34模型和数据增强的深度伪造视频检测方法 Deepfake Videos Detection Method Based on i_ResNet34 Model and Data Augmentation 计算机科学, 2021, 48(7): 77-85. https://doi.org/10.11896/jsjkx.210300258 |
|
||
首页