A visual data completion method based on low-rank tensor estimation defining a nuclear norm

Abstract

The invention belongs to the technical field of image or video data calculation methods, and particularly relates to a visual data completion method based on low-rank tensor estimation of a limited nuclear norm. The method comprises the following steps of A1, initializing a pre-prepared original tensor to be complemented to obtain a target tensor x, wherein the original tensor to be complemented comprises a color image and a video sequence; A2, constructing a limited kernel norm model for tensor completion by using the target tensor x to obtain a tensor completion target function; A3, performing optimization modeling on the objective function to obtain an alternative function capable of solving the optimal solution of the objective function; and A4, solving the substitution function to obtain a target tensor, converting the target tensor into a data source corresponding format, and obtaining a final completion result of the color image and the video sequence. According to the method, an improved restriction nuclear norm model is extended to tensor data completion to approach the rank of original tensor data, so that tensor data completion is completed.

Description

technical field [0001] The invention belongs to the technical field of image or video data calculation methods, and in particular relates to a visual data completion method based on low-rank tensor estimation of a limited nuclear norm. Background technique [0002] In computer vision research, many visual data such as color images and video sequences often have very complex high-level data structures, and traditional data representations such as vectors and matrices cannot well reflect the internal structure of these multi-dimensional data. [0003] As a generalization of higher order (order greater than or equal to 3) represented by vector (first order) and matrix (second order), tensor can better express the essential structure of multidimensional data such as images and videos. In the process of actually obtaining data, some elements in the data may be lost due to transmission, encoding and data conversion. The technique of predicting and recovering unknown elements from...

AI Technical Summary

Problems solved by technology

If overlapping tensors cannot be estimated correctly in practical applications, the final completion result, such as the completion of color images, will cause problems such as blurring of image edge details

In addition, the model of the patented completion method is the sum of the trace norms of all decomposed small tensors. This model itself cannot well approximate the rank of the original tensor, which will also lead to a poor final completion effect.

Images