Motion Key-frames extraction based on amplitude of distance characteristic curve

摘要: AbstractThe key frames extraction technique extracts postures to describe the original motion sequence, which has been widely used in compression, retrieval, edition and so on. In this paper, we propose a method based on amplitude of curve find captured sequence. First select group joint distance features represent adopt Principal Component Analysis (PCA) obtain one dimension principal component as will be used. Then gain initial key-frames by extracting local optimum points curve. At last, get final inserting merging too close. A number experimental examples demonstrate that our is practicable efficient not only visual performance but also aspect compression ratio error rate.