REGIMRobvid: Objects and Scenes Detection for Robot Vision 2013.

摘要

Abstract. This paper describes the participation of the REGIM team in the Im-ageCLEF 2013 Robot Vision Challenge. The competition was focused on the problem of objects and scenes classification in indoor environments. Objects and scenes are considered as concepts. During the competition, we aim to classify images according to the room in which they were acquired, using the informa-tion provided by the visual images only. Our system is based on PHOW features extraction and PEGASOS SVM algorithm to learn a multi-class classifier that is enable to detect the objects and the adequate scene. For this end, we focus on how to interpret the scores provided by the SVM classifier. Our system was ranked 4th among 6 teams.