Image dark data value evaluation method based on similarity hash

Abstract

The invention discloses an image dark data value evaluation method based on similarity hash. The method comprises the following steps: obtaining dark data sets, inputting the dark data set into a trained DSTH model; obtaining a hash code of each image in the dark data set; constructing a hash map according to the hash code of each image in the obtained dark data set, wherein a node in the hash mapis a hash code of the image, the connection edge between the nodes is an element in the constructed adjacent matrix; obtaining the importance score of each node in the obtained hash map, sorting theimportance scores of all the obtained nodes according to a descending order, and outputting an image corresponding to the importance score of the first k bits in a sorting result to a user. Aiming atthe problems that enterprises occupy huge storage space for owned large-scale image dark data and are lack of management and utilization, the storage cost performance of the image dark data with neglected value is improved by 'lightening ' the image dark data.

Description

technical field [0001] The invention belongs to the technical field of computer image mining, and more particularly relates to a method for evaluating the value of image dark data based on similarity hashing. Background technique [0002] With the rapid development of cloud computing and the Internet, the exponential growth of data has brought huge costs to storage systems and data centers. However, as an unlabeled and unrelated data resource, dark data has always occupied a large amount of storage space, but it is difficult to exert its own value at present. Contents of the invention [0003] Aiming at the above defects or improvement needs of the prior art, the present invention provides a method and system for evaluating the value of image dark data based on similarity hash. The problem of storage space and lack of management and utilization, by "lighting up" the image dark data whose value is ignored, improves its storage cost performance. [0004] In order to achiev...

AI Technical Summary

Problems solved by technology

However, as an unlabeled and unrelated data resource, dark data has always occupied a large amount of storage space, but it is difficult to exert its own value at present

Images