ID 原文 译文
52987 其凭借自身出色的性能被学术界和工业界高度认可。 With its excellent performance, it is highly recognized by the academic and industrial circles.
52988 本文主要研究NVDLA在FPGA平台上的优化映射问题， This article mainly studies the optimization mapping problem of NVDLA on the FPGA platform.
52989 通过多种优化方案高效利用FPGA内部的硬件资源，同时提高其运行性能。 Through various optimization schemes, the hardware resources inside the FPGA are efficiently utilized and the running performance is improved.
52990 基于搭建的NVDLA加速器平台，本文实现了对RESNET-50神经网络的硬件加速，完成了在ImageNet数据集上的图像分类任务。 Based on the built NVDLA accelerator platform, hardware acceleration of the RESNET-50 neural network is achieved, and the image classification task on the ImageNet dataset is completed.
52991 研究结果表明，优化后的NVDLA能显著提高硬件资源使用效率，处理性能最高可达30. 8 fps，实现了较边缘中央处理器(CPU)加速器平台28倍的性能提升。 The test results show that the optimized NVDLA can significantly improve the utilization efficiency of hardware resources, and the processing performance can reach up to 30. 8 fps, which is a 28 times performance improvement over the edge central processing unit(CPU) accelerator platform.
52992 针对基于区域的目标检测算法中定位精度不高的问题，本文提出了一种分裂机制的改进Faster RCNN算法。 For the region-based target detection algorithm, there is a general problem that the boundary position is far from the real value. This paper proposes an improved Faster RCNN algorithm using the splitting mechanism.
52993 该算法首先选择特征提取能力强的卷积神经网络(CNN)作为骨干网络进行特征的提取; 然后通过12种不同Anchors产生候选目标区，以进一步提升检测的精确度; Firstly, the algorithm selects the convolutional neural network(CNN) with strong feature extraction ability as the backbone network to extract features, and then generates candidate target regions through 12 different anchors to further improve the accuracy of detection.
52994 最后将得到的特征分别传送到两个子网络，分别实现分类与定位。 Finally, the obtained features are transmitted to two different sub-networks:the classification network is based on the fully-connected structure, and the targets are classified;
52995 分类网络以全连接结构为基础，定位网络则主要由卷积神经网络构成。 the positioning network is based on the convolutional neural network structure to achieve the target positioning.
52996 本文在Pascal VOC2007和Pascal VOC2012以及吸尘袋图像集上对算法的有效性进行了验证。 The experiments verify the effectiveness of the algorithm on the Pascal VOC2007 dataset, Pascal VOC2012 dataset and vacuum bag dataset.