3 results about "Kernel clustering" patented technology

The application discloses a multi-scale structure preserving optimal transmission depth kernel clustering method, and belongs to the technical field of unsupervised machine learning, pattern recognition and data mining, and comprises the following steps: S1, multi-layer coding is performed on input unlabeled single-view data to obtain hierarchical hidden representation; S2, each layer of normalized kernel matrix is obtained based on the step S1; S3, a fusion kernel matrix is obtained based on the step S2; S4, the fusion kernel matrix obtained based on the step S2; S5, a multi-scale structure preserving regularization term of the step S3 is taken as a core constraint; and S6, after the final fusion kernel matrix is obtained after model training convergence, a multi-scale structure preserving regularization is constructed from the local neighborhood distribution consistency and the global geometric correlation, the geometric deviation of the kernel space relative to the input space is explicitly constrained, and the excessive geometric distortion and structure degradation in the kernel learning process are inhibited.

The invention discloses a method and system for identifying various foreign matters based on deep learning, and relates to the technical field of image identification, and the method comprises the steps: conveying a material through a conveyor belt, and dynamically adjusting the illumination gear according to the speed of the conveyor belt in an image collection process, so as to enable the same region to obtain an image sample under a multi-illumination condition; performing feature identification on each frame of foreign matter by using a feature extraction network, dividing feature subsets according to illumination gears, and identifying samples with incomplete features through consistency discrimination; an open type authigenic nucleus clustering structure is adopted, and a primary nucleus formed by a standard sample and an authigenic nucleus formed by a sample with incomplete characteristics jointly participate in clustering; performing category inference on the authigenic nucleus through the radiation effect of the native nucleus on the feature dimension and multi-scale pyramid generality analysis; and finally, obtaining a final recognition result of the foreign matter based on multi-frame feature weighted fusion. According to the method, high-robustness foreign matter recognition under multi-illumination, multi-scale and multi-frame fusion is realized.

The invention discloses a program code optimization method and device in a link stage, equipment and a storage medium, and the method comprises the steps: traversing each target file input into a linker, and obtaining a program section in each target file; if the program section is a clustering base point function, creating a kernel clustering instance corresponding to the clustering base point function, and executing clustering optimization operation on the clustering base point function according to the kernel clustering instance; and after the clustering base point functions in all the target files complete the clustering optimization operation, executing a linking operation on each target file. According to the technical scheme, the feasibility of the code optimization method can be improved, the influence of instruction cache deletion is eliminated, and the performance of an executable program is greatly improved.