A simple seed growing algorithm for estimating scene flow in a stereo setup is presented. Two calibrated and synchronized cameras observe a scene and output a sequence of image pairs. The algorithm simultaneously computes a disparity map between the image pairs and optical flow maps between consecutive images. This, together with calibration data, is an equivalent representation of the 3D scene flow, i.e. a 3D velocity vector is associated with each reconstructed point. The proposed method starts from correspondence seeds and propagates these correspondences to their neighborhood. It is accurate for complex scenes with large motions and produces temporallycoherent stereo disparity and optical flow results. The algorithm is fast due to inherent search space reduction. An explicit comparison with recent methods of spatiotemporal stereo and variational optical and scene flow is provided.

No abstract

No abstract

Abstract-In this paper we present a method for detecting and localizing an active speaker, i.e., a speaker that emits a sound, through the fusion between visual reconstruction with a stereoscopic camera pair and sound-source localization with several microphones. Both the cameras and the microphones are embedded into the head of a humanoid robot. The proposed statistical fusion model associates 3D faces of potential speakers with 2D sound directions. The paper has two contributions: (i) a method that discretizes the two-dimensional space of all possible sound directions and that accumulates evidence for each direction by estimating the time difference of arrival (TDOA) over all the microphone pairs, such that all the microphones are used simultaneously and symmetrically and (ii) an audio-visual alignment method that maps 3D visual features onto 2D sound directions and onto TDOAs between microphone pairs. This allows to implicitly represent both sensing modalities into a common audiovisual coordinate frame. Using simulated as well as real data, we quantitatively assess the robustness of the method against noise and reverberations, and we compare it with several other methods. Finally, we describe a realtime implementation using the proposed technique and with a humanoid head embedding four microphones and two cameras: this enables natural human-robot interactive behavior.

No abstract

Abstract. An action recognition algorithm which works with binocular videos is presented. The proposed method uses standard bag-of-words approach, where each action clip is represented as a histogram of visual words. However, instead of using classical monocular HoG/HoF features, we construct features from the scene-flow computed by a matching algorithm on the sequence of stereo images. The resulting algorithm has a comparable or slightly better recognition accuracy than standard monocular solution in controlled setup with a single actor present in the scene. However, we show its significantly improved performance in case of strong background clutter due to other people freely moving behind the actor.

Abstract-In this paper we address the problem of audiovisual speaker detection. We introduce an online system working on the humanoid robot NAO. The scene is perceived with two cameras and two microphones. A multimodal Gaussian mixture model (mGMM) fuses the information extracted from the auditory and visual sensors and detects the most probable audio-visual object, e.g., a person emitting a sound, in the 3D space. The system is implemented on top of a platformindependent middleware and it is able to process the information online (17Hz). A detailed description of the system and its implementation are provided, with special emphasis on the online processing issues and the proposed solutions. Experimental validation, performed with five different scenarios, show that that the proposed method opens the door to robust humanrobot interaction scenarios.

This paper shows that Hidden Markov Models (HMMs) can be effectively applied to 3D face data. The examined HMM techniques are shown to be superior to a previously examined Gaussian Mixture Model (GMM) technique. Experiments conducted on the Face Recognition Grand Challenge database show that the Equal Error Rate can be reduced from 0.88% for the GMM technique to 0.36% for the best HMM approach.