Research On Image Feature Extraction Algorithm Based On LML Model

Image feature extraction is the premise of image recognition and analysis.Whether the extracted feature information can accurately characterize the image determines the subsequent image analysis work.When performing image feature extraction,directly performing feature extraction on an image will occupy a large storage space.Therefore,it is necessary to find a method that can not only ensure the main information of the image but also reduce the dimension of the image.Based on the Lie group machine learning,the unitary group,a typical Lie group,is introduced to reduce the dimension of highdimensional images.The algebraic representation is used to digitize the image,and a mathematical description is given to the feature dimension reduction of the image.In addition,the unitary group machine learning model is applied to image data dimensionality reduction,and the image features after dimensionality reduction are extracted.The main work includes the following aspects:1)Aiming at the problem that the image samples occupy a large storage space,the key is to reduce the dimensionality of the image,and choose to use the outer product of vectors to define the basic image and express the image in the form of a matrix.2)Perform singular value decomposition on the image matrix,decompose to obtain the unitary matrix representing the image,and use the unitary matrix to form a unitary group.Based on the Lie group machine learning model,the unitary group machine learning model is constructed by combining the Lie group’s characteristics about groups and manifolds.3)Taking the leaf image data as the experimental data,the image is decomposed by singular value.When selecting different numbers of singular values,the leaf images with different effects are obtained.It is found that the image with 100 singular values is enough to represent the original image,so as to reduce the dimension of image data and greatly reduce the number of bits characterizing image features.Finally,the corner feature of the leaf image processed by the unitary group machine learning model is extracted.The research results show that the number of feature points extracted from the leaf image processed by the unitary group machine learning model reaches 150,which is higher than that of other corner detection algorithms,and the distribution of leaf corner points is more uniform.After verifying that the unitary group machine learning model reduces the dimensionality of the image data,the accuracy and superiority of the corner feature extraction algorithm are guaranteed.It expands the application of Lie group machine learning in image feature extraction and promotes the intelligent development of image recognition field.Figure 30;Table 2;Reference 51...