.png){kind=icon}
.png){kind=icon}
Return
Context-based camera selection from multiple video streams
.png){kind=icon}
PRE
AI
DOI：10.1007/s11042-021-11674-6.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Multi-camera systems are often used for broadcasting various events such as sports, education, concerts or meetings. In such systems, automatic editing is necessary to choose the camera to broadcast according to the main action in the scene. However, automatic camera selection methods often are specific to a particular event. In this work, we propose to model the functionalities of automatic editing systems based on the analysis of the state of the art in order to isolate the application context. This allows us to propose an automatic camera selection method based on a knowledge-model. This model aims to represent the specification and formalisation of the knowledge of an application context. The proposed method has been successfully applied for municipal councils broadcast. This leads to create a context event ontology - Person Of Interest (POI), Action Of Interest (AOI) - related to municipal council. Moreover, in order to offer live broadcasting, a new speaker detection method is proposed. The speaker detection method has been checked on a large ground truth with an accuracy better than 98%.
Keywords:
Knowledge modelling
Camera selection
Automatic editing
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3
Papers:
1.9W
Citations:
3.2W