Beauty Pageant of Koi Method Based on Improved ResNeXt50 Residual Network
-
摘要: 锦鲤选美的不同等级之间具有高相似度的特点，目前都是人工进行选美分级。为解决人工选美所存在的效率低、主观性强、成本高的问题，提出了一种基于迁移学习和改进ResNeXt50残差网络的锦鲤选美方法。本文首先构建了红白、大正、昭和3种锦鲤的选美等级数据集。其次，采用迁移学习策略提高训练速度，并从SE注意力模块、Hardswish激活函数和Ranger优化器3方面对ResNeXt50模型进行了改进，构建了SH-ResNeXt50锦鲤选美分级模型。试验结果表明：SH-ResNeXt50模型有效提升了锦鲤选美的等级分选能力，模型准确率达95.6%,损失值仅0.074,优于常用的AlexNet、GoogLeNet、ResNet50和ResNeXt50网络模型。最后，采用Grad-CAM分析SH-ResNeXt50模型的可解释性，结果表明SH-ResNeXt50模型和人工识别的感兴趣区域基本一致。本文所提出的方法实现了具有高相似度的锦鲤不同等级的智能分选，对其它具有高相似度的生物等级分选具有借鉴意义。Abstract: There is a high similarity among different levels of the koi for beauty pageant, and beauty grading for koi is currently done manually. To solve these problems of low efficiency, strong subjectivity and high cost of manual beauty pageants, a sorting method for koi beauty pageant was proposed based on transfer learning and improved ResNeXt50 residual network. Firstly, a rank dataset was constructed for the beauty pageant on Kohaku, Taisho and Showa koi. Secondly, the transfer learning strategy was adopted to improve the training speed and improve the ResNeXt50 model from three aspects of SE attention module, Hardswish activation function and Ranger optimizer, further a SH-ResNeXt50 classification model was proposed and constructed for koi pageant. The experimental results showed that the SH-ResNeXt50 model effectively improved the sorting ability for koi beauty pageant, with an accuracy of 95.6% and a loss value of only 0.074, which was better than the commonly used AlexNet, GoogLeNet, ResNet50 and ResNeXt50 network models. Finally, the interpretability of SH-ResNeXt50 model was analyzed by Grad-CAM, and the results showed that the regions of interest of SH-ResNeXt50 model was basically consistent with those recognized by the humans. The approach proposed realized the intelligent sorting of different levels of koi beauty pageant with high similarity, which had reference significance for other biological level sorting with high similarity.
-
-
[1] 李潇轩,李志辉,尹陆乐,等.浅谈锦鲤文化传承与发展[J].水产养殖,2020,41(3):78-80.LI Xiaoxuan,LI Zhihui,YIN Lule,et al.Discussion on the cultural inheritance and development of koi carp[J].Aquaculture,2020,41(3):78-80.(in Chinese) [2] BAIRWA M K,SAHARAN N,RAWAT K D,et al.Effect of LED light spectra on reproductive performance of koi carp (Cyprinus carpio)[J].Indian Journal of Animal Research,2017,51(6):1012-1018.
[3] 彭芳,陈康勇,钟为铭,等.三种锦鲤的分类和鉴赏[J].科学养鱼,2018(8):82-83.PENG Fang,CHEN Kangyong,ZHONG Weiming,et al.Classification and appreciation of three species of koi[J].Sci.Fish Farming,2018(8):82-83.(in Chinese) [4] 宋述芹,段培旭.锦鲤及其品种分类[J].齐鲁渔业,2009(26):53-54.SONG Shuqin,DUAN Peixu.Koi and its variety classification[J].Shandong Fish,2009(26):53-54.(in Chinese) [5] 李道亮,王帅星,王聪.柔性可穿戴传感技术在智慧渔业中的应用进展[J].农业工程学报,2023,39(13):1-13.LI Daoliang,WANG Shuaixing,WANG Cong.Application of flexible wearable sensing technology in smart fishery[J].Transactions of the CSAE,2023,39(13):1-13.(in Chinese) [6] 赵博,宋正河,毛文华,等.基于PSO与K-均值算法的农业超绿图像分割方法[J].农业机械学报,2009,40(8):166-169.ZHAO Bo,SONG Zhenghe,MAO Wenhua,et al.Agriculture extra-green image segmentation based on PSO and K-means clustering[J].Transactions of the Chinese Society for Agricultural Machinery,2009,40(8):166-169.(in Chinese) [7] GREWAL P S,OLOUMIl F,RUBIN U,et al.Deep learning in ophthalmology:a review[J].Canadian Journal of Ophthalmology-Journal Canadien D Ophtalmologie,2018,53(4):309-313.
[8] LIU P,ZHANG C F,QI H,et al.Multi-attention DenseNet:a scattering medium imaging optimization framework for visual data pre-processing of autonomous driving systems[J].IEEE Transactions on Intelligent Transportation Systems,2022,23(12):396-407.
[9] 戚超,左毅,陈哲琪,等.基于改进VGG16的大米加工精度分级方法研究[J].农业机械学报,2021,52(5):301-307.QI Chao,ZUO Yi,CHEN Zheqi,et al.Rice processing accuracy classification method based on improved VGG16 convolution neural network[J].Transactions of the Chinese Society for Agricultural Machinery,2021,52(5):301-307.(in Chinese) [10] 高新怡,陈琦,陈冠宇,等.卷积神经网络在实时检测领域的研究[J].软件工程,2022,25(6):22-29.GAO Xinyi,CHEN Qi,CHEN Guanyu,et al.Research on convolutional neural networks in real-time detection[J].Software Engineering,2022,25(6):22-29.(in Chinese) [11] 万鹏,赵竣威,朱明,等.基于改进ResNet50模型的大宗淡水鱼种类识别方法[J].农业工程学报,2021,37(12):159-168.WAN Peng,ZHAO Junwei,ZHU Ming,et al.Freshwater fish species identification method based on improved ResNet50 model[J].Transactions of the CSAE,2021,37(12):159-168.(in Chinese) [12] GONG B,DAI K Y,SHAO J,et al.Fish-TViT:a novel fish species classification method in multi water areas based on transfer learning and vision transformer[J].Heliyon,2023,9(6):e16761.
[13] ZHENG Y L,DENG L M,LIN Q,et al.KRS-Net:a classification approach based on deep learning for koi with high similarity[J].Biology-Basel,2022,11(12):1727.
[14] JU Z Y,XUE Y J.Fish species recognition using an improved AlexNet model[J].Optik,2020,223:165499.
[15] HUANG X,CHEN W,YANG W.Improved algorithm based on the deep integration of googlenet and residual neural network[J].Journal of Physics:Conference Series,2021,1757(1):012069.
[16] HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),2016:770-778.
[17] XIE S,GIRSHICK R,DOLLAR P,et al.Aggregated residual transformations for deep neural networks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),2017:5987-5995.
[18] 袁培森,宋进,徐焕良.基于残差网络和小样本学习的鱼图像识别[J].农业机械学报,2022,53(2):282-290.YUAN Peisen,SONG Jin,XU Huanliang.Fish image recognition based on residual network and few-shot learning[J].Transactions of the Chinese Society for Agricultural Machinery,2022,53(2):282-290.(in Chinese) [19] 彭红星,何慧君,高宗梅,等.基于改进ShuffleNetV2模型的荔枝病虫害识别方法[J].农业机械学报,2022,53(12):290-300.PENG Hongxing,HE Huijun,GAO Zongmei,et al.Litchi diseses and insect pests identification method improved ShuffleNetV2[J].Transactions of the Chinese Society for Agricultural Machinery,2022,53(12):290-300.(in Chinese) [20] 孙龙清,王新龙,王泊宁,等.基于ResNet-CA的鱼群饱腹程度识别方法[J].农业机械学报,2022,53(增刊2):219-225,277.SUN Longqing,WANG Xinlong,WANG Boning,et al.Identification method of fish satiation level based on ResNet-CA[J].Transactions of the Chinese Society for Agricultural Machinery,2022,53(Supp.2):219-225,277.(in Chinese) [21] CHEN J D,ZHANG D,ZEB A,et al.Identification of rice plant diseases using lightweight attention networks[J].Expert Systems With Applications,2021,169:114514.
[22] AVILES J C,KOUKI A.Position-aided mm-wave beam training under NLOS conditions[J].IEEE Access,2016,4:8703-8714.
[23] HU J,SHEN L,SUN G.Squeeze-and-Excitation networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition,2018:7132-7141.
[24] 薛勇,王立扬,张瑜,等.基于GoogLeNet深度迁移学习的苹果缺陷检测方法[J].农业机械学报,2020,51(7):30-35.XUE Yong,WANG Liyang,ZHANG Yu,et al.Defect detection method of apples based on GoogLeNet deep transfer learning[J].Transactions of the Chinese Society for Agricultural Machinery,2020,51(7):30-35.(in Chinese) [25] 徐建鹏,王杰,徐祥,等.基于RAdam卷积神经网络的水稻生育期图像识别[J].农业工程学报,2021,37(8):143-150.XU Jianpeng,WANG Jie,XU Xiang,et al.Image recognition for different developmental stages of rice by RAdam deep convolutional neural networks[J].Transactions of the CSAE,2021,37(8):143-150.(in Chinese) [26] SELVARAJU R R,COGSWELL M,DAS A,et al.Grad-CAM:visual explanations from deep networks via gradient-based localization[J].International Journal of Computer Vision,2020,128(2):336-359.
计量
- 文章访问数: 0
- HTML全文浏览量: 0
- PDF下载量: 0
下载:
