针对遥感图像中目标方向、尺度和长宽比变化剧烈以及时而密集时而稀疏排列等造成的特征提取困难和基于回归的旋转检测器中存在的边界不连续问题,我们提出了以下解决方案:首先,使用YOLOv7检测器进行特征提炼与融合;其次,结合密集编码标签...针对遥感图像中目标方向、尺度和长宽比变化剧烈以及时而密集时而稀疏排列等造成的特征提取困难和基于回归的旋转检测器中存在的边界不连续问题,我们提出了以下解决方案:首先,使用YOLOv7检测器进行特征提炼与融合;其次,结合密集编码标签(Densely Coded Label,DCL),构建角度分类编码-解码器,实现旋转检测;然后,改进SimOTA自适应样本匹配策略为R-SimOTA,通过在代价函数中增加角度分类损失指导,提高样本分配准确性;最后,在损失函数中增加角度分类损失(DCL Loss)和权值(theta),并引入角度距离和长宽比敏感加权(Angle Distance and Aspect Ratio Sensitive Weighting,ADARSW),指导模型收敛并拟合出正确的角度预测值。在大型遥感图像数据集DOTA上进行了大量实验和视觉分析,结果表明该方案有效。展开更多
A compact antenna formed by three concentric split rings for ultra-high frequency(UHF)radio frequency identification(RFID)tag is proposed in this paper.The antenna is composed of two parts,an outer short-circuited rin...A compact antenna formed by three concentric split rings for ultra-high frequency(UHF)radio frequency identification(RFID)tag is proposed in this paper.The antenna is composed of two parts,an outer short-circuited ring modified from a traditional split-ring resonator(SRR)antenna and an inner SRR load,so the antenna can be regarded as a short-circuited ring loaded with SRR.According to the transmission line theory,to conjugate match with the capacitive input-impedance of a tag chip,the length of the short-circuited ring isλg/4 shorter than that of an open-circuited dipole of a traditional SRR antenna,whereλg is the wavelengh of the operating frequency.Hence,the size of the proposed antenna is more compact than that of the traditional SRR antenna.Thereafter,the proposed antenna is simulated and optimized by ANSYS high-frequency structure simulator(HFSS).The impedance,efficiency,and mutual coupling of the fabricated antenna are tested in a reverberation chamber(RC).The results show that the size of the presented antenna is 83%smaller than that of the traditional SRR antenna and the proposed antenna can cover the whole UHF RFID operating frequency band worldwide(840—960 MHz).The measured read range of the tag exhibits maximum values of 45 cm in free space and 37 cm under dense tag environment.展开更多
Accurate head poses are useful for many face-related tasks such as face recognition, gaze estimation,and emotion analysis. Most existing methods estimate head poses that are included in the training data(i.e.,previous...Accurate head poses are useful for many face-related tasks such as face recognition, gaze estimation,and emotion analysis. Most existing methods estimate head poses that are included in the training data(i.e.,previously seen head poses). To predict head poses that are not seen in the training data, some regression-based methods have been proposed. However, they focus on estimating continuous head pose angles, and thus do not systematically evaluate the performance on predicting unseen head poses. In this paper, we use a dense multivariate label distribution(MLD) to represent the pose angle of a face image. By incorporating both seen and unseen pose angles into MLD, the head pose predictor can estimate unseen head poses with an accuracy comparable to that of estimating seen head poses. On the Pointing'04 database, the mean absolute errors of results for yaw and pitch are 4.01?and 2.13?, respectively. In addition, experiments on the CAS-PEAL and CMU Multi-PIE databases show that the proposed dense MLD-based head pose estimation method can obtain the state-of-the-art performance when compared to some existing methods.展开更多
文摘针对遥感图像中目标方向、尺度和长宽比变化剧烈以及时而密集时而稀疏排列等造成的特征提取困难和基于回归的旋转检测器中存在的边界不连续问题,我们提出了以下解决方案:首先,使用YOLOv7检测器进行特征提炼与融合;其次,结合密集编码标签(Densely Coded Label,DCL),构建角度分类编码-解码器,实现旋转检测;然后,改进SimOTA自适应样本匹配策略为R-SimOTA,通过在代价函数中增加角度分类损失指导,提高样本分配准确性;最后,在损失函数中增加角度分类损失(DCL Loss)和权值(theta),并引入角度距离和长宽比敏感加权(Angle Distance and Aspect Ratio Sensitive Weighting,ADARSW),指导模型收敛并拟合出正确的角度预测值。在大型遥感图像数据集DOTA上进行了大量实验和视觉分析,结果表明该方案有效。
文摘A compact antenna formed by three concentric split rings for ultra-high frequency(UHF)radio frequency identification(RFID)tag is proposed in this paper.The antenna is composed of two parts,an outer short-circuited ring modified from a traditional split-ring resonator(SRR)antenna and an inner SRR load,so the antenna can be regarded as a short-circuited ring loaded with SRR.According to the transmission line theory,to conjugate match with the capacitive input-impedance of a tag chip,the length of the short-circuited ring isλg/4 shorter than that of an open-circuited dipole of a traditional SRR antenna,whereλg is the wavelengh of the operating frequency.Hence,the size of the proposed antenna is more compact than that of the traditional SRR antenna.Thereafter,the proposed antenna is simulated and optimized by ANSYS high-frequency structure simulator(HFSS).The impedance,efficiency,and mutual coupling of the fabricated antenna are tested in a reverberation chamber(RC).The results show that the size of the presented antenna is 83%smaller than that of the traditional SRR antenna and the proposed antenna can cover the whole UHF RFID operating frequency band worldwide(840—960 MHz).The measured read range of the tag exhibits maximum values of 45 cm in free space and 37 cm under dense tag environment.
基金supported by the National Key Scientific Instrument and Equipment Development Project of China(No.2013YQ49087903)the National Natural Science Foundation of China(No.61202160)
文摘Accurate head poses are useful for many face-related tasks such as face recognition, gaze estimation,and emotion analysis. Most existing methods estimate head poses that are included in the training data(i.e.,previously seen head poses). To predict head poses that are not seen in the training data, some regression-based methods have been proposed. However, they focus on estimating continuous head pose angles, and thus do not systematically evaluate the performance on predicting unseen head poses. In this paper, we use a dense multivariate label distribution(MLD) to represent the pose angle of a face image. By incorporating both seen and unseen pose angles into MLD, the head pose predictor can estimate unseen head poses with an accuracy comparable to that of estimating seen head poses. On the Pointing'04 database, the mean absolute errors of results for yaw and pitch are 4.01?and 2.13?, respectively. In addition, experiments on the CAS-PEAL and CMU Multi-PIE databases show that the proposed dense MLD-based head pose estimation method can obtain the state-of-the-art performance when compared to some existing methods.
