Automatic Video Segmentation through Editing Analysis

J. M. Corridoni?, A. Del Bimbo?y

? Dipartimento di Sistemi e Informatica, Universit?a degli Studi di Firenze,

via S. Marta 3, 50139 Firenze, Italy.
y Dipartimento di Elettronica per l'Automazione, Universit?a degli Studi di Brescia,

via Branze, 38, 25133 Brescia, Italy

SUMMARY

In the present paper the problem of video segmentation for indexing and retrieval in multimedia databases is addressed. Large archives of video sequences will remain a largely unexploitable resource of information, unless a straightforward way to describe their content is designed. The first step in this direction is stating that a film, or generally speaking a video, is the product of the editing of a set of shots, which can be considered as the syntactic units of the visual speech. Therefore, video segmentation must focus on the editing procedures, that generate cuts, dissolves, fades, through which the shots are fused into the video stream. The parsing of the continuous video data into these units is herein addressed. This work proposes a statistically based new algorithm to detect cuts between two shots in a video sequence. Such algorithm is data driven, since no a priori knowledge about the shots is available. On the contrary, the analysis of how the other editing techniques are performed, supplies a mathematical model, upon which the algorithms for the detection of fades dissolves and mattes are based. All the algorithms are inserted into a global framework, in order to generate a low level parsing and segmentation of video streams.

Keywords: Video{Databases Video{Segmentation Video{Indexing Video Editing Process.

Addressed Topics

1. Image languages and DBMS.

2. Application domains.