Cross-domain video semantic concept detection method based on active learning

Abstract

The invention relates to a cross-domain video semantic concept detection method based on active learning, comprising the following steps of: (1) taking a Gaussian random field as a standard classifier; (2) selecting and marking unmarked samples by adopting an uncertainty query strategy in an active learning method; (3) updating the standard classifier; and (4) sequentially repeating the steps (2) and (3) until a certain cycle index is completed. In the invention, the Gaussian random field is used for constructing the standard classifier for active learning. Compared with the way that only an original domain sample is taken as a training set of the standard classifier, which is frequently used in other active learning algorithms, the standard classifier in the invention has the advantage that the selected marked samples can reflect distribution of data in a target domain to a greater extent. Weight of a newly selected marked sample from the target domain is increased, thus the standard classifier can rapidly adapt to characteristic distribution of the data in the target domain. The invention also provides an algorithm for rapidly updating a standard model, the complexity of the algorithm is effectively reduced, and the applicability of the algorithm is improved.

Description

technical field [0001] The invention relates to a cross-domain video semantic concept detection method based on active learning, which belongs to the field of video content analysis and semantic concept detection. Background technique [0002] Video semantic concept detection is to automatically detect semantic concepts appearing in videos, such as "car", "person" and "building". However, with the increase of massive video data, more and more videos come from different fields, such as news videos, network videos and document videos. Since the same semantic concept is distributed differently in the feature space of videos from different domains, when a semantic concept classifier using video data from one domain as the training set is tested on video data from another domain, it will get poor classification Effect. The simple solution is to train a new classifier for each semantic concept of a new domain. However, due to the large number of semantic concepts contained in th...

AI Technical Summary

Problems solved by technology

However, the traditional active learning method assumes that the test set and the training set have the same data distribution. However, in the cross-domain video semantic concept detection, the video data of the new target domain is different from the original domain data in the feature space distribution.

Images