Person Identity Label Propagation in Stereo Videos

Abstract

Abstract-In this paper a novel method is introduced for propagating person identity labels on facial images extracted from stereo videos. It operates on image data with multiple representations and calculates a projection matrix that preserves locality information and a priori pairwise information, in the form of must-link and cannot-link constraints between the various data representations. The final data representation is a linear combination of the projections of all data representations. Moreover, the proposed method takes into account information obtained through data clustering. This information is exploited during the data propagation step in two ways: to regulate the similarity strength between the projected data and to indicate which samples should be selected for label propagation initialization. The performance of the proposed Multiple Locality Preserving Projections with Cluster-based Label Propagation (MLPP-CLP) method was evaluated on facial images extracted from stereo movies. Experimental results showed that the proposed method outperforms state of the art methods.

Multimodal Stereoscopic Movie Summarization Conforming to Narrative Characteristics

“…As in the case of speakers, each face appearance consists simply of a video segment that starts and ends at the temporal boundaries of an uninterrupted face appearance. Such data may have been acquired through the successive application of face detection [52], face tracking [53], face clustering [54] and label propagation [55] algorithms.…”

Feature extraction and statistical analysis of videos for cinemetric applications

“…All information was extracted from the proposed framework. Here we should note that we applied a semi-automatic annotation procedure for each extracted facial image, giving actor labels by employing a label propagation algorithm [16]. That is, we employed a semi-supervised approach where we have manually annotated 10% of the exacted facial images and the remaining were automatically annotated.…”

Movie shot selection preserving narrative properties

“…As in the case of speakers, each face appearance consists simply of a video segment that starts and ends at the temporal boundaries of an uninterrupted face appearance. Such data may have been acquired through the successive application of face detection [12], face tracking [13], face clustering [14] and label propagation [15] algorithms. Despite these algorithmic prerequisites, no extra data modalities (such as the movie script) are required, beyond the film itself.…”