Key frame extraction method based on cosine distance hierarchical clustering

Abstract

The invention discloses a key frame extraction method based on cosine distance hierarchical clustering, so as to reuse existing action data and reduce redundant information in action capturing data. The rotation amount of a joint point is used as a feature value for segmenting the motion capturing data, noise in the capturing data is removed, high-dimensional data are mapped to low-dimensional data through a dimension reduction method, a cosine distance is then used to calculate a similarity, hierarchical clustering is then adopted for segmentation, and a frame with the minimum Euclidean distance error between the frame gesture in each segmentation point and each section and the mean value is used as a key frame sequence. Extraction on the key frame is high in accuracy and a recall ratio, the extracted key frame has strong summarization and presentation skills, and the method can be used for segmenting the high-dimensional motion capturing data.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a key frame extraction method based on cosine distance hierarchical clustering. Background technique [0002] 3D animation synthesis technology is a new technology developed on the basis of motion capture technology. Since its creation, this technology has been widely used in digital media fields such as game animation, movie special effects, and virtual reality. In recent years, with the steady improvement of computer performance and the breakthrough and precision improvement of recording equipment and motion capture sensor technology, 3D animation synthesis technology has become a research hotspot in the field of computer graphics. At present, the realization basis of this technology still depends on the motion capture technology. Motion capture technology is the premise of this technology. The data information of the motion to be captured is obtained through the sens...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a key frame extraction method based on cosine distance hierarchical clustering, which solves the redundancy of a large amount of captured data and the problem of segmenting high-dimensional data, so as to improve the utilization rate of captured data and reduce the high sampling rate of motion capture equipment influence, thereby improving the accuracy of target detection

Images