Geometric skewimage classification algorithm based on fusion adaptive reasoning

摘要

The geometric skew model is widely used in image classification and image retrieval. In the traditionalskew model, the geometric topology statistical method ignores the spatial informationbetween geometricwords and the classification object shape information, which leads to the distinguishingability of image features. In this paper, we propose a geometric skew method withadaptive reasoning. Combined with significant region extraction and geometric topology, we notonly can generate more representative geometric topologies but also can avoid complex backgroundinformation and position change bands to a certain extent. First, the adaptive reasoninggeometric skewness model is constructed on the obtained significant area by carrying out the significantregion extraction of the training image. Second, in order to describe the characteristics ofthe image more accurately, it can resist the changing position and background information. Themethod uses the geometric topology strategy and the polygon segmentation method to integrateglobal information and local information. Through the simulation experiment and compared withthe traditional skewness model and other models, the results show that the method proposed inthis paper obtains higher intelligence image classification accuracy.