Scenario image annotation method based on active learning and multi-label multi-instance learning

Abstract

The present invention is directed to two fundamental characteristics of a scene image: (1) the scene image often containing complex semantics; and (2) a great number of manual annotation images taking high labor cost. The invention further discloses a scene image annotation method based on an active learning and a multi-label and multi-instance learning. The method comprises: training an initial classification model on the basis of a label image; predicting a label to an unlabeled image; calculating a confidence of the classification model; selecting an unlabeled image with the greatest uncertainty; experts carrying on a manual annotation on the image; updating an image set; and stopping when an algorithm meeting the requirements. An active learning strategy utilized by the method ensures accuracy of the classification model, and significantly reduces the quantity of the scenario image needed to be manually annotated, thereby decreasing the annotation cost. Moreover, according to the method, the image is converted to a multi-label and multi-instance data, complex semantics of the image has a reasonable demonstration, and accuracy of image annotation is improved.

Description

technical field [0001] The invention relates to the technical field of scene image labeling, in particular to a scene image labeling method based on active learning and multi-label multi-instance learning. Background technique: [0002] With the development of information technology and the advancement of Internet services, various websites such as news, social networking, and commodity trading have been greatly developed, and the Internet generates a large number of scene pictures every day. These scene pictures have the following two basic characteristics. On the one hand, a single scene image not only reflects one content, but may involve multiple topics, and the semantics are relatively complex. For example, an image about a street may involve many different subjects such as pedestrians, roads, vehicles, trees, sky, buildings, etc. [0003] On the other hand, a large number of scene images generated by the Internet do not have classification labels that can fully desc...

AI Technical Summary

Problems solved by technology

[0005]The purpose of this invention is to solve the two basic characteristics of the scene image, the scene image may contain multiple content areas, the semantics are complex, and it is impossible to convert it into a single vector accurately A scene image annotation method based on multi-instance multi-label learning and active learning to represent the theme of the scene image, and a large number of scene images on the Internet do not have classification labels, and the annotation cost is expensive.

Images