Detecting keypoints in dairy cows aims to locate and track the motion trajectories of the body's joints,which plays a crucial role in behavior analysis and lameness detection.However,real farming scenarios,charact...Detecting keypoints in dairy cows aims to locate and track the motion trajectories of the body's joints,which plays a crucial role in behavior analysis and lameness detection.However,real farming scenarios,characterized by occlusions and large variations in object scale may result in poor detection results.Therefore,we introduce the atrous spatial pyramid pooling(ASPP) module into the shallow layers network of ResNet101,designed to improve the multi-scale feature extraction capability of the model.The ASPP module enhances the robustness of recognition for different dimensional sizes and occluded keypoints using different dilatation rates in the parallel atrous convolutional layers to expand the model's receptive field.Furthermore,seven types of motion features,including tracking up,gait symmetry,step height balance,motion speed variability,head swing amplitude,head-neck slope and back curvature are extracted simultaneously by monitoring and tracking the motion trajectory of distinct keypoints.Several of these features represent innovative extraction models and attributes,first proposed in this study.Multiple models are trained and tested on datasets containing 2,385 frames for ablation experiments.The experiments show that,in comparison with the ResNet50,MobileNet_v2_1.0,and EfficientNet-b0backbone networks,the training error and test error of ResNet101 are reduced by 4.04-30.12 pixels and 3.81-28.14 pixels.Therefore,ResNet101 is used as the benchmark for subsequent model improvement by adding the ASPP module.The training error and test error of the ResNet101-ASPP network are reduced by 0.27 and 0.24 pixels,respectively,compared to the benchmark network.The prediction confidence improves by 1.65-2.50% at three different dairy cow object scales.In addition,the keypoints under different occlusion conditions improve considerably,especially for small-scale keypoints,demonstrating the capability of the ASPP module for multi-scale feature extraction.By analyzing the distribution of the seven features and health,mild lameness,and severe lameness in dairy cows,it is shown that all the different features play an important role in distinguishing between different levels of lameness.展开更多
【目的】针对滚动轴承工作环境恶劣、早期退化趋势微弱,导致预测过程的故障特征提取难度大、预测精度不高等问题,提出一种基于累积变换的强化诊断模型,对滚动轴承开展剩余使用寿命预测任务。【方法】首先,提出了基于累积变换的特征强化...【目的】针对滚动轴承工作环境恶劣、早期退化趋势微弱,导致预测过程的故障特征提取难度大、预测精度不高等问题,提出一种基于累积变换的强化诊断模型,对滚动轴承开展剩余使用寿命预测任务。【方法】首先,提出了基于累积变换的特征强化方法,将提取所得特征量转换为对应的累积变换形式,增强原始特征量的敏感度;其次,构建基于累积特征的新型健康指标,利用连续触发机制算法对健康指标进行状态划分,获得初始故障发生点;最后,强化ResNet的跳连接模块、增加额外的校准通道,提高网络模型对关键退化特征的聚焦能力,并通过带Kolmogorov-Arnold Network(KAN)模块的堆叠长短期记忆(Stacked Long Short-Term Memory with KAN Module,KSLSTM)网络,获取全时空特征,进行轴承剩余使用寿命的精准预测。【结果】结果表明,通过构建累积变换特征和优化网络结构来提高小样本训练环境下的轴承剩余使用寿命预测精度,仿真和试验证明了该方法的有效性。展开更多
Quantum error correction technology is based on the principle of redundant encoding,encoding logical quantum information into multiple physical qubits to provide important support for the stable operation of quantum c...Quantum error correction technology is based on the principle of redundant encoding,encoding logical quantum information into multiple physical qubits to provide important support for the stable operation of quantum computers.To address the issues of low decoding accuracy and limited feature extraction in quantum error correction,this paper proposes a toric code decoder based on a syndrome-preliminary error fusion module(SPEFM)and a ResNet architecture.This decoder takes full advantage of the correlations between X and Z errors.In the SPEFM,the syndrome and preliminary error predictions are deeply fused,while a unidirectional Swin transformer architecture is incorporated to extract global error features from the syndrome data,signiffiificantly improving both decoding accuracy and computational efffiificiency.In addition,this paper further extracts local error features from the fused features using the deep residual structure of ResNet,enhancing the decoder's ability to capture quantum error patterns.Experimental results show that the decoder is applicable to different code distances(d=4,6,8,10)under the depolarizing noise model.Its bit error rate is lower than that of the minimum weight perfect matching(MWPM)algorithm,and its logical error rate is lower than both the MWPM algorithm and the ResNet18 decoder.Furthermore,the decoding threshold is increased to 0.163,representing a 3.82%improvement over the MWPM algorithm threshold of 0.157.展开更多
电子元器件种类繁多且没有一致的细粒度分类标准,为快速满足元器件在不同粒度下的分类需求,提出一种基于深度学习的YOLOR-ECA(YOLOv8 and ResNet50 with efficient channel attention)电子元器件检测算法。首先采用YOLOv8网络定位元器...电子元器件种类繁多且没有一致的细粒度分类标准,为快速满足元器件在不同粒度下的分类需求,提出一种基于深度学习的YOLOR-ECA(YOLOv8 and ResNet50 with efficient channel attention)电子元器件检测算法。首先采用YOLOv8网络定位元器件位置,然后采用ResNet50网络对定位获取的元器件进行识别分类,通过元器件种类的增减满足不同细粒度的分类标准。为提升模型对尺寸小、特征相似元器件的细节特征提取能力,分类网络引入ECA注意力机制,并对残差结构的捷径连接部分进行改进;为避免神经元失活,采用GELU(Gaussian Error Linear Units)激活函数。实验结果表明,改进的YOLOR-ECA模型的检测准确率为96.6%,并且对于小尺寸元器件的识别精度最高可达100%,对于具有特征相似性元器件的误检率最低可降到0.01%,能实现电子元器件在不同细粒度分类标准下的高效检测。展开更多
基金supported by the National Natural Science Foundation of China (32102600)the Central Publicinterest Scientific Institution Basal Research Fund, China (Y2023XK13, JBYW-AII-2024-28/40, and JBYWAII-2023-33/37/42)+1 种基金Science and Technology Innovation Project of the Chinese Academy of Agricultural Sciences (CAAS-ASTIP-2021-AII)the Wuhu Science and Technology Bureau Two Strong One Increase Project, China (2023ly12)。
摘要Detecting keypoints in dairy cows aims to locate and track the motion trajectories of the body's joints,which plays a crucial role in behavior analysis and lameness detection.However,real farming scenarios,characterized by occlusions and large variations in object scale may result in poor detection results.Therefore,we introduce the atrous spatial pyramid pooling(ASPP) module into the shallow layers network of ResNet101,designed to improve the multi-scale feature extraction capability of the model.The ASPP module enhances the robustness of recognition for different dimensional sizes and occluded keypoints using different dilatation rates in the parallel atrous convolutional layers to expand the model's receptive field.Furthermore,seven types of motion features,including tracking up,gait symmetry,step height balance,motion speed variability,head swing amplitude,head-neck slope and back curvature are extracted simultaneously by monitoring and tracking the motion trajectory of distinct keypoints.Several of these features represent innovative extraction models and attributes,first proposed in this study.Multiple models are trained and tested on datasets containing 2,385 frames for ablation experiments.The experiments show that,in comparison with the ResNet50,MobileNet_v2_1.0,and EfficientNet-b0backbone networks,the training error and test error of ResNet101 are reduced by 4.04-30.12 pixels and 3.81-28.14 pixels.Therefore,ResNet101 is used as the benchmark for subsequent model improvement by adding the ASPP module.The training error and test error of the ResNet101-ASPP network are reduced by 0.27 and 0.24 pixels,respectively,compared to the benchmark network.The prediction confidence improves by 1.65-2.50% at three different dairy cow object scales.In addition,the keypoints under different occlusion conditions improve considerably,especially for small-scale keypoints,demonstrating the capability of the ASPP module for multi-scale feature extraction.By analyzing the distribution of the seven features and health,mild lameness,and severe lameness in dairy cows,it is shown that all the different features play an important role in distinguishing between different levels of lameness.
摘要【目的】针对滚动轴承工作环境恶劣、早期退化趋势微弱,导致预测过程的故障特征提取难度大、预测精度不高等问题,提出一种基于累积变换的强化诊断模型,对滚动轴承开展剩余使用寿命预测任务。【方法】首先,提出了基于累积变换的特征强化方法,将提取所得特征量转换为对应的累积变换形式,增强原始特征量的敏感度;其次,构建基于累积特征的新型健康指标,利用连续触发机制算法对健康指标进行状态划分,获得初始故障发生点;最后,强化ResNet的跳连接模块、增加额外的校准通道,提高网络模型对关键退化特征的聚焦能力,并通过带Kolmogorov-Arnold Network(KAN)模块的堆叠长短期记忆(Stacked Long Short-Term Memory with KAN Module,KSLSTM)网络,获取全时空特征,进行轴承剩余使用寿命的精准预测。【结果】结果表明,通过构建累积变换特征和优化网络结构来提高小样本训练环境下的轴承剩余使用寿命预测精度,仿真和试验证明了该方法的有效性。
基金supported by the Joint Fund of the Natural Science Foundation of Shandong Province,China(Grant Nos.ZR2022LLZ012 and ZR2021LLZ001)the Key Research and Development Program of Shandong Province,China(Grant No.2023CXGC010901)。
摘要Quantum error correction technology is based on the principle of redundant encoding,encoding logical quantum information into multiple physical qubits to provide important support for the stable operation of quantum computers.To address the issues of low decoding accuracy and limited feature extraction in quantum error correction,this paper proposes a toric code decoder based on a syndrome-preliminary error fusion module(SPEFM)and a ResNet architecture.This decoder takes full advantage of the correlations between X and Z errors.In the SPEFM,the syndrome and preliminary error predictions are deeply fused,while a unidirectional Swin transformer architecture is incorporated to extract global error features from the syndrome data,signiffiificantly improving both decoding accuracy and computational efffiificiency.In addition,this paper further extracts local error features from the fused features using the deep residual structure of ResNet,enhancing the decoder's ability to capture quantum error patterns.Experimental results show that the decoder is applicable to different code distances(d=4,6,8,10)under the depolarizing noise model.Its bit error rate is lower than that of the minimum weight perfect matching(MWPM)algorithm,and its logical error rate is lower than both the MWPM algorithm and the ResNet18 decoder.Furthermore,the decoding threshold is increased to 0.163,representing a 3.82%improvement over the MWPM algorithm threshold of 0.157.
摘要电子元器件种类繁多且没有一致的细粒度分类标准,为快速满足元器件在不同粒度下的分类需求,提出一种基于深度学习的YOLOR-ECA(YOLOv8 and ResNet50 with efficient channel attention)电子元器件检测算法。首先采用YOLOv8网络定位元器件位置,然后采用ResNet50网络对定位获取的元器件进行识别分类,通过元器件种类的增减满足不同细粒度的分类标准。为提升模型对尺寸小、特征相似元器件的细节特征提取能力,分类网络引入ECA注意力机制,并对残差结构的捷径连接部分进行改进;为避免神经元失活,采用GELU(Gaussian Error Linear Units)激活函数。实验结果表明,改进的YOLOR-ECA模型的检测准确率为96.6%,并且对于小尺寸元器件的识别精度最高可达100%,对于具有特征相似性元器件的误检率最低可降到0.01%,能实现电子元器件在不同细粒度分类标准下的高效检测。