为解决从图像的低层视觉特征到高层语义特征的"语义鸿沟"问题,对当前的语义提取方法进行研究,简单介绍了图像语义层次模型,并根据语义信息的来源不同,归纳总结了图像语义中基于处理范围的方法,基于机器学习的方法,基于人机交互的方法和基于外部信息源的提取方法,这些工作为图像语义提取和图像语义检索等研究提供有益参考。
The current image semantic extraction method is researched to find a solution to eliminate the "semantic gap" between low-level visual features and high-level semantic features of images.The image semantic level model is simply introduced.According to the semantic information extracted from different sources,the information extraction methods based on processing region,machine learning,man-machine alternation and external information source are summed up.The above work provides a valuable reference for image semantic extraction and retrieval.