A crowd gathering sensitive image detection method

Abstract

The invention discloses a crowd gathering sensitive image detection method, which comprises the following steps: collecting image data in the gathering crowd, respectively obtaining image data sets ofthe original domain and the target domain; establishing a supervised discriminant sparse transfer model, which can learn the common subspace and the sparse representation matrix of the original domain and the target domain simultaneously, and adding a supervised discriminant regular term to the transfer model; obtaining the common subspace of the classification task in the original domain and thetarget domain by using the label information of the image samples, and implementing the optimization process by transfer learning in the common subspace; using the alternating optimization strategy and ADMM algorithm to solve the optimization variables alternately; recognizing sensitive images. A supervise discrimination sparse migration model is established to improve that correct rate of classification task under the condition of small sample, thereby greatly improving the correct rate and discrimination of sensitive image recognition in complex environment such as crowd gathering and thelike.

Description

technical field [0001] The invention belongs to the technical field of image detection, in particular to a crowd gathering sensitive image detection method. Background technique [0002] With the rapid growth of data set size and computing resources, the theory and application of artificial intelligence and its subfield machine learning have made great progress. Especially in the field of computer vision, the exponential explosion of data set size is particularly obvious. In real life, a large number of images are generated every day, but there are very few image data with label information, because manual labeling data consumes a huge amount of time and money. At the same time, in the classification of crowd-sensitive images, the tasks are changeable, and the scene and lighting changes greatly, making classification more difficult. [0003] In the framework of traditional machine learning, whether it is a classification or regression task, a mapping function or a classifi...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above problems, the present invention proposes a crowd-gathering-sensitive image detection method. Aiming at problems such as the scarcity of label samples in the target domain and the inconsistency of features, the establishment of a supervised discriminant sparse transfer model is used to improve the classification task in the case of small samples. accuracy, thus greatly improving the accuracy and discrimination of sensitive image recognition in complex environments such as crowd gathering

Images