Information retrieval method

Abstract

The invention provides an information retrieval method which comprises the following steps: extracting the edge of a video image by using an adaptive edge detection algorithm to obtain a binary edge image; blocking the video area, and calculating the proportion of white pixels in the current block; calculating the number of black blocks obtained after blocking; if the number of the black blocks is greater than a preset threshold, extracting a video target for retrieving by a video segmenting method based on edge features, and comparing the Euclidean distance between every two images. According to the information retrieval method provided by the invention, a better retrieval result for the video images of which the target and the background are distinguished is realized, and the accuracy is improved.

Description

technical field [0001] The invention relates to video image retrieval, in particular to an information retrieval method. Background technique [0002] Content-based video retrieval achieves the purpose of retrieval by extracting video image features and performing feature matching. Most of the existing technologies are retrieval based on underlying features. Since the underlying features are difficult to express people's subjective concept of the target in the video, and the process of human beings being able to recognize the meaning of the video requires a lot of life experience and reasoning, there is still a large fault zone between the underlying features and high-level semantic features of the video. Improving video retrieval accuracy and retrieval speed is not enough to popularize and apply. For example, different videos have different characteristics, so if only a certain characteristic of the video is used for feature extraction, the accuracy of the retrieval resul...

AI Technical Summary

Problems solved by technology

Because the underlying features are difficult to express people's subjective concept of the target in the video, and the process of human beings being able to recognize the meaning of the video requires a lot of life experience and reasoning, so there are still large fault zones between the underlying features and high-level semantic features of videos. Improving the accuracy and speed of video retrieval is not enough to promote the application

For example, different videos have different characteristics, so if only a certain characteristic of the video is used for feature extraction, the accuracy of the retrieval results will be affected

[0003] Therefore, for the above-mentioned problems existing in related technologies, effective solutions have not yet been proposed

Images