计算机科学 ›› 2024, Vol. 51 ›› Issue (11A): 230900037-7.doi: 10.11896/jsjkx.230900037
张雷1, 武文喆1, 白雪媛2
ZHANG Lei1, WU Wenzhe1, BAI Xueyuan2
摘要： 为实现围棋对弈过程中的高精度实时记谱,提出了一种基于结合三维注意力机制与轻量化卷积的实时检测算法Light-YOLOv8。在YOLOv8模型的基础上,使用PWConv+PConv替换主干网络中跨阶段局部网络的3*3卷积,大幅减少模型计算量与参数规模;加入CARAFE上采样算子与SimAM三维注意力机制,提高对围棋目标的检测能力;使用Wise-IOU损失函数提高模型定位能力与收敛速度,提高了对棋子粘连、棋子重叠与光照不均匀情况下的检测能力。在自定义围棋数据集上进行对比训练表明,改进后的算法实现了检测精度的提升与推理速度的提高。针对移动端设备部署需求对模型进行优化与压缩,并在不同安卓设备部署,图像分辨率为640*480的情况下,结合图像预处理与后处理操作,拍照检测平均时间为89 ms,平均模型推理帧率为37.6 fps。进行50轮记谱实验,平均记谱准确率高于97%,平均胜负判别准确率到达100%,能够实现稳定的围棋记谱功能。
中图分类号:
| [1]LIANG C Y,DING Y,YUAN J,et al.Research on Image Re-cognition Algorithms for Go Tournaments [J].Modern Compu-ter(Professional Edition),2016(24):39-43. [2]MAO L M,ZHU P Y,LU Z L,et al.Recognition of Gobang Robot Chessboard Based on LabVIEW [J].Computer Engineering and Design,2017,38(1):242 246. [3]GUI Y,WU Y,WANG Y,et al.Visual Image Processing of Humanoid Go Game Robot Based on OPENCV[C]//2020 Chinese Control And Decision Conference(CCDC).Hefei,2020:3713-3716. [4]WANG Y J,ZHANG Y B,WU Y Y,et al.A General Chess Po-sitioning Method under Uneven Illumination [J].Computer Applications,2020,40(12):3490-3498. [5]ZHANG Y,GU F.A Compact and Low-cost gobangg Robot[C]//2022 International Conference on Service Robotics(ICoSR).Chengdu,2022:62-66. [6]LIU W,ANGUELOV D,ERHAN D,et al.Ssd:Single shotmultibox detector[C]//Computer Vision－ECCV 2016:14th European Conference,Amsterdam,The Netherlands,Part I 14.Springer International Publishing,2016:21-37. [7]WU B,IANDOLA F,JINP H,et al.Squeezedet:Unified,small,low power fully convolutional neural networks for real-time object detection for autonomous driving[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops.2017:129-137. [8]TAN M,LE Q.Efficientnetv2:Smaller models and faster trai-ning[C]//International Conference on Machine Learning.PMLR,2021:10096-10106. [9]JIANG P,ERGU D,LIU F,et al.A Review of Yolo algorithm developments[J].Procedia Computer Science,2022,199:1066-1073. [10]GIRSHICK R.Fast r-cnn[C]//Proceedings of the IEEE International Conference on Computer Vision.2015:1440-1448. [11]HE K,GKIOXARI G,DOLLÁR P,et al.Mask r-cnn[C]//Proceedings of the IEEE International Conference on Computer Vision.2017:2961-2969. [12]PURKAITP,ZHAO C,ZACH C.SPP-Net:Deep absolute pose regression with synthetic views[J].arXiv:1712.03452,2017. [13]KONG F G,LI Z Z,LIU Q,et al.Design of the Vision System for Chinese Chess Robots [J].Mechanical Engineer,2020,347(5):1-3. [14]LIU X,JIN J,ZHONG Q,et al.Vision-Based Chess Detection for a Robotic Companion[C]//2021 IEEE 6th International Conference on Computer and Communication Systems(ICCCS).Chengdu,2021:248-253. [15]HOWARDA G,ZHU M,CHEN B,et al.Mobilenets:Efficient convolutional neural networks for mobile vision applications[J].arXiv:1704.04861,2017. [16]FANG G F,MA X Y,SONG M L,et al.[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).2023:16091-16101. [17]CHEN J,KAO S,HE H,et al.Run,don't walk:chasing higher FLOPS for faster neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2023:12021-12031. [18]WANG J,CHEN K,XU R,et al.CARAFE:Content-aware reassembly of features[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.2019:3007-3016. [19]YANG L,ZHANG R Y,LI L,et al.SimAM:A simple,parameter-free attention module for convolutional neural networks[C]//International Conference on Machine Learning.PMLR,2021:11863-11874. [20]TONG Z,CHEN Y,XU Z,et al.Wise-IoU:bounding box re-gression loss with dynamic focusing mechanism[J].arXiv:2301.10051,2023. |
|
||
首页