ID	原文	译文
57298	实验结果表明，所提算法在织锦作品风格迁移任务中的表现优于现有算法.	It is shown that the proposed approach generates brocade stylization outputs that have high quality as compared with other ap- proaches.
57299	在多样化信息传输的通信系统中，若无差别地传输具有不同统计特性与不同重要性的信源，将带来能量资源的浪费.	In the mobile communication system with diversified information transmission，if the sources with different statistical characteristics and different importance are transmitted without difference，the en- ergy resources would be wasted.
57300	鉴于此，以信源特性为依据，提出了一种结合联合编码与中继译码转发的不等错误保护编译码方案.	Therefore，based on the characteristics of the source，an unequal error protection scheme combining joint source channel coding in relay system is proposed.
57301	在发送端，针对经离散余弦变换后图像信源的统计特性和重要性程度采取相应的传输策略:	The proposed scheme adopts different transmission strategies according to the source statistical characteristics and the importance:
57302	对熵值较小且重要程度不高的高频信息进行信源压缩，提高传输效率;对重要程度较高的低频信息利用中继进行二次信道编码，实现不等保护.	the high frequency information with low entropy and less importance is compressed to improve the transmission efficiency，and the low frequency information with more importance is coded once more in relay to provide more protection.
57303	在接收端，设计多模块联合迭代译码方式，充分利用不同传输策略下不同形式的信源信息，提升系统性能.	At the receiver，a joint decoding method including multiple module itera- tions is designed to improve system performance by making full use of the different representation of the source information under different transmission strategies.
57304	理论分析和仿真结果表明，对比传统方案，所提方案在相同的峰值信噪比下，有 1 dB 以上的信噪比增益.	Analysis and simulation show that compared with the traditional schemes，the proposed scheme has more than 1 dB performance gain under the same peak signal-to-ratio.
57305	针对目前数据标注过于依赖硬件、手动数据标注效率低下的问题，提出了基于深度学习的人体图像半自动标注系统.	In view of the problem that data labeling is too dependent on hardware and manual data labe- ling is inefficient，a semi-automatic labeling system for human images based on deep learning is proposed.
57306	系统通过对算法进行改进，增加人体关键点个数进行特征提取和加入运动信息的约束，提高了视频分阶段标注的准确率.	By improving the algorithm，the system increases the number of key points of the human body for feature extraction and adds motion information constraints，which improves the accuracy of video staged annota- tion.
57307	使用真实数据集仿真实验证明了通过深度学习算法进行数据标注的可行性，并且使用半自动标注的速度快、准确率高.	Experiments that employs real data sets prove the feasibility of data labeling by deep learning algo- rithm，and using deep learning algorithms for semi-automatic labeling is faster and more accurate.