计算机科学 ›› 2024, Vol. 51 ›› Issue (4): 229-235.doi: 10.11896/jsjkx.230100137
宋昊, 毛宽民, 朱洲
SONG Hao, MAO Kuanmin, ZHU Zhou
摘要: 端到端的立体匹配算法在计算时间和匹配效果上均有一定的优势,近年来在立体匹配任务中得到了广泛的应用。但特征提取的过程中存在特征冗余、信息丢失,以及多尺度特征融合不充分等问题,造成算法的计算量和复杂度偏高,也影响了匹配的精度。针对上述问题,在自适应聚合网络AANET的基础上,设计了更加适合立体匹配的特征提取模块,提出了改进的幽灵自适应聚合网络GAANET。采用G-Ghost阶段提取多尺度的特征,通过廉价操作生成部分特征,减少特征的冗余现象并有效保存浅层特征;采取高效的通道注意力机制,将不同的权重分配到每个通道中;采取改进的特征金字塔结构,缓解传统金字塔中的通道信息丢失并优化融合特征,为各个尺度的特征进行丰富的信息补充。在SceneFlow,KITTI2015和KITTI2012数据集上进行训练和评估,评估结果显示,与基础方法相比,所提改进算法的精度分别提升了0.92%,0.25%和0.20%,且参数量减少了13.75%,计算量减少了4.8%。
中图分类号:
[1]LIN X,WANG J,LIN C.Research on 3D Reconstruction in Bi-nocular Stereo Vision Based on Feature Point Matching Method[C]//2020 IEEE 3rd International Conference on Information Systems and Computer Aided Education(ICISCAE).Dalian:IEEE,2020:551-556. [2]CHEN Y,ZHAO L W,ZHAN H C,et al.Study on Reconstruction of Indoor 3D Scene Based on Binocular Vision[J].Computer Science,2020,47(11A):175-177. [3]CHANG Z T,SHI Y Q,WANG J,et al.Vehicle Speed Measure-ment Method Based on Binocular Vision[J].Computer Science,2021,48(9):135-139. [4]DE SANTANA J R,CAMBUIM L F S,BARROS E.Bi-Window Based Stereo Matching Using Combined Siamese Convolutional Neural Network[C]//13th International Conference on Digital Image Processing(ICDIP).Electr Network:SPIE,2021:118781Z. [5]SEKI A,POLLEFEYS M.SGM-Nets:Semi-global matchingwith neural networks[C]//30th IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu:IEEE,2017:6640-6649. [6]JIE Z Q,WANG P F,LING Y G,et al.Left-Right Comparative Recurrent Model for Stereo Matching[C]//31st IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Salt Lake City:IEEE,2018:3838-3846. [7]WANG Y,LAI Z H,HUANG G,et al.Anytime Stereo Image Depth Estimation on Mobile Devices[C]//IEEE International Conference on Robotics and Automation(ICRA).Montreal:IEEE,2019:5893-5900. [8]KHAMIS S,FANELLO S,RHEMANN C,et al.Stereonet:Guided hierarchical refinement for real-time edge-aware depth prediction[C]//Proceedings of the European Conference on Computer Vision(ECCV).2018:573-590. [9]LIU S G,ZHANG T,YANG J G,et al.Progressive dialtion residual network for deep binocular stereo matching[J].Journal of Xidian University,2022,49(5):175-180. [10]LI T,MA W,XU S B,et al.Task-Adaptive End-to-End Net-works for Stereo Matching[J].Journal of Computer Research and Development,2020,57(7):1531-1538. [11]HAN K,WANG Y,XU C,et al.GhostNets on Heterogeneous Devices via Cheap Operations[J].International Journal of Computer Vision,2022,130(4):1050-1069. [12]WANG Q,WU B,ZHU P,et al.ECA-Net:Efficient Channel Attention for Deep Convolutional Neural Networks[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Seattle:IEEE,2020:11534-11542. [13]LUO Y,CAO X,ZHANG J,et al.CE-FPN:enhancing channel information for object detection[J].Multimedia Tools and Applications,2022,81(21):30685-30704. [14]XU H F,ZHANG J Y.AANet:Adaptive Aggregation Network for Efficient Stereo Matching[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Electr Network:IEEE,2020:1956-1965. [15]CHABRA R,STRAUB J,SWEENEY C,et al.StereoDRNet:Dilated Residual StereoNet[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Long Beach:IEEE,2019:11778-11787. [16]DAI J,QI H,XIONG Y,et al.Deformable Convolutional Networks[C]//2017 IEEE International Conference on Computer Vision(ICCV).Venice:IEEE,2017:764-773. [17]CHENG X,ZHONG Y,HARANDI M,et al.Hierarchical neural architecture search for deep stereo matching[J].Advances in Neural Information Processing Systems,2020,33:22158-22169. [18]SHEN Z,DAI Y C,SONG X B,et al.PCW-Net:Pyramid Combination and Warping Cost Volume for Stereo Matching[C]//17th European Conference on Computer Vision(ECCV).Tel Aviv:Springer,2022:280-297. [19]SHEN Z L,DAI Y C,RAO Z B,et al.CFNet:Cascade and Fused Cost Volume for Robust Stereo Matching[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Electr Network:IEEE,2021:13901-13910. [20]WANG Q,SHI S H,ZHENG S Z,et al.FADNet:A Fast and Accurate Network for Disparity Estimation[C]//IEEE International Conference on Robotics and Automation(ICRA).Electr Network:IEEE,2020:101-107. |
|