ID 原文 译文
52877 最后，针对构建的全向移动机器人进行自主导航实验，实验结果验证了所设计系统的实用性。 Finally, the autonomous navigation experiment is carried out for the omnidirectional mobile robot, and the experimental results verify the practicability of the designed system.
52878 目标检测是计算机视觉任务中的关键问题之一，而边界框回归是目标检测的重要步骤。 Target detection is one of the key problems in computer vision tasks, and bounding box regression is an important step in target detection.
52879 现有方法认为交并比(IoU)损失有利于度量IoU，但仍然存在收敛速度慢和回归不准确的问题。 In the existing methods, intersection over union(IoU) loss is considered to be beneficial to IoU measurement, but there are still problems such as slow convergence and inaccurate regression.
52880 此外，大规模的目标检测数据集尽可能地清晰定义真实边界框，但在标记边界框时仍然会引入偏差。 In addition, large-scale target detection data sets try to define the true bounding box as clearly as possible, but biases are still introduced when marking the bounding box.
52881 本文基于CIoU损失和边界框回归Kullback-Leibler(KL)损失提出了优化IoU(OIoU)边界框损失， Therefore, this paper proposes an optimized bounding box optimize-IoU(OIoU) loss based on complete-IoU(CIoU) loss and bounding box regression Kullback-Leibler(KL)loss,
52882 主要通过学习边界变换和定位方差解决上述问题。 and solves the above problems mainly by learning boundary transformation and positioning variance.
52883 OIoU不仅用于评估本地化置信度还可以计算标准差和边界框位置，在解决边界框标注时带来的偏差、回归不准确问题的同时加快了收敛速度。 OIoU is not only used to evaluate the localization confidence, but also to calculate the standard deviation and the position of the bounding box, which solves the problems of deviation and inaccurate regression caused by the bounding box labeling while speeding up the convergence speed.
52884 在PASCAL VOC和MS-COCO两个数据集上证明了优化边界框损失在检测精度和收敛速度上要优于原有的方法。 On the two data sets of PASCAL VOC and MS-COCO, it is proved that the optimized bounding box loss is superior to the previous methods in detection accuracy and convergence speed.
52885 全封闭往复式压缩机作为制冷系统的核心部件，其品质决定了整机系统的能效水平、静音效果和产品寿命。 As the core component of the refrigeration system, the quality of hermetic reciprocating compressor determines the energy efficiency level, silent effect, and product life of the whole system.
52886 在生产线制造过程中，针对全封闭结构特点难以识别缺陷产品的短板问题，本文提出了一种基于壳体振动信号的压缩机制造缺陷诊断方法。 In the manufacturing process of the production line, in order to solve the shortcomings that it is difficult to identify the defective product due to the characteristics of the hermetic structure, this paper proposes a diagnosis method for manufacturing defects based on the vibration signal of the compressor shell.