Sumarização automática de rushes vídeos baseada em características espaciais e espaço-temporais

Abstract

Advances in compression techniques, the continuous decreasing on equipment acquisition costs and video storage and availability of high speed data transmission, facilitates the way videos are created, stored and distributed. As a result, videos are now used in various applications, which made data management tasks complex and inefficient. To help solving these problems, automatic video summarization techniques emerged and have been attracting research interest in recent years. Video summarization, as the name implies, is a brief summary of the content of a video. Specifically, a video summary is a sequence of still or moving images that represent the content of a video.This work presents an approach to solving the problem of automatic generation of dynamic video summaries. The method benefits from the use of unsupervised classification, spatial and spatiotemporal features for summaries generation. The evaluation is made by users and the results are compared with works found in the literature. The proposed methodology achieved statistically similar quality in relation to the approaches compared.