基于深度语义的人体部位分割算法研究
Algorithm for Deep Semantic Segmentation of Human Body Parts
尹红梅 1金格格 2华继镛3
作者信息
- 1. 扬州海关,江苏扬州 225009
- 2. 扬州技师学院,江苏扬州 225300
- 3. 扬州大学,江苏扬州 225000
- 折叠
摘要
人体可以划分成多个部位,各部位在空间上存在语义关系.在自建的人体图像集上,通过标注工具,按12个目标类别标注训练样本.在传统语义模型基础上,改进了数据清洗方法,增加了模型评估.实验证明,改进后的深度语义分割算法具有良好效果.
Abstract
The human body can be divided into multiple parts,and each part has semantic relationships in space.We annotated training samples based on 12 target categories using annotation tools on our self-built hu-man body image set.Based on traditional semantic models,data cleaning methods have been improved and model evaluation has been added.Experimental results have shown that the improved deep semantic segmenta-tion algorithm has good performance.
关键词
深度学习/语义分割/图像处理/目标分割Key words
deep learning/semantic segmentation/image processing/object segmentation引用本文复制引用
出版年
2024
NETL
NSTL
万方数据