The analysis of video acquired with a wearable camera is a challenge that multimedia community is facing with the proliferation\r\nof such sensors in various applications. In this paper, we focus on the problem of automatic visual place recognition in a weakly\r\nconstrained environment, targeting the indexing of video streams by topological place recognition.We propose to combine several\r\nmachine learning approaches in a time regularized framework for image-based place recognition indoors.Theframework combines\r\nthe power of multiple visual cues and integrates the temporal continuity information of video. We extend it with computationally\r\nefficient semisupervised method leveraging unlabeled video sequences for an improved indexing performance. The proposed\r\napproach was applied on challenging video corpora. Experiments on a public and a real-world video sequence databases show\r\nthe gain brought by the different stages of the method.
Loading....