We introduce a distinctive real-time, causal, neural network-based active speaker detection system optimized for low-power edge computing. This system drives a virtual cinematography module and is deployed on a commercial device. The system uses data originating from a microphone array and a 360-degree camera. Our network requires only 127 MFLOPs per participant, for a meeting with 14 participants. Unlike previous work, we examine the error rate of our network when the computational budget is exhausted, and find that it exhibits graceful degradation, allowing the system to operate reasonably well even in this case. Departing from conventional DOA estimation approaches, our network learns to query the available acoustic data, considering the detected head locations. We train and evaluate our algorithm on a realistic meetings dataset featuring up to 14 participants in the same meeting, overlapped speech, and other challenging scenarios.

This paper describes a system that generates speaker-annotated transcripts of meetings by using a microphone array and a 360-degree camera. The hallmark of the system is its ability to handle overlapped speech, which has been an unsolved problem in realistic settings for over a decade. We show that this problem can be addressed by using a continuous speech separation approach. In addition, we describe an online audio-visual speaker diarization method that leverages face tracking and identification, sound source localization, speaker identification, and, if available, prior speaker information for robustness to various real world challenges. All components are integrated in a meeting transcription framework called SRD, which stands for “separate, recognize, and diarize”. Experimental results using recordings of natural meetings involving up to 11 attendees are reported. The continuous speech separation improves a word error rate (WER) by 16.1% compared with a highly tuned beamformer. When a complete list of meeting attendees is available, the discrepancy between WER and speaker-attributed WER is only 1.0%, indicating accurate word-to-speaker association. This increases marginally to 1.6% when 50% of the attendees are unknown to the system.

Development of a rich hand-gesture-based interface is currently a tedious process, requiring expertise in computer vision and/or machine learning. We address this problem by introducing a simple language for pose and gesture description, a set of development tools for using it, and an algorithmic pipeline that recognizes it with high accuracy. The language is based on a small set of basic propositions, obtained by applying four predicate types to the fingers and to palm center: direction, relative location, finger touching and finger folding state. This enables easy development of a gesture-based interface, using coding constructs, gesture definition files or an editing GUI. The language is recognized from 3D camera input with an algorithmic pipeline composed of multiple classification/regression stages, trained on a large annotated dataset. Our experimental results indicate that the pipeline enables successful gesture recognition with a very low computational load, thus enabling a gesture-based interface on low-end processors.

We present a new real-time hand tracking system based on a single depth camera. The system can accurately reconstruct complex hand poses across a variety of subjects. It also allows for robust tracking, rapidly recovering from any temporary failures. Most uniquely, our tracker is highly flexible, dramatically improving upon previous approaches which have focused on front-facing close-range scenarios. This flexibility opens up new possibilities for human-computer interaction with examples including tracking at distances from tens of centimeters through to several meters (for controlling the TV at a distance), supporting tracking using a moving depth camera (for mobile scenarios), and arbitrary camera placements (for VR headsets). These features are achieved through a new pipeline that combines a multi-layered discriminative reinitialization strategy for per-frame pose estimation, followed by a generative model-fitting stage. We provide extensive technical details and a detailed qualitative and quantitative analysis.

We present the Discriminative Ferns Ensemble (DFE) classifier for efficient visual object recognition. The classifier architecture is designed to optimize both classification speed and accuracy when a large training set is available. Speed is obtained using simple binary features and direct indexing into a set of tables, and accuracy by using a large capacity model and careful discriminative optimization. The proposed framework is applied to the problem of hand pose recognition in depth and infra-red images, using a very large training set. Both the accuracy and the classification time obtained are considerably superior to relevant competing methods, allowing one to reach accuracy targets with run times orders of magnitude faster than the competition. We show empirically that using DFE, we can significantly reduce classification time by increasing training sample size for a fixed target accuracy. Finally a DFE result is shown for the MNIST dataset, showing the method's merit extends beyond depth images.

Unsupervised feature learning is the task of using unlabeled examples for building a representation of objects as vectors. This task has been extensively studied in recent years, mainly in the context of unsupervised pre-training of neural networks. Recently, (Coates et al., 2011) conducted extensive experiments, comparing the accuracy of a linear classifier that has been trained using features learnt by several unsupervised feature learning methods. Surprisingly, the best performing method was the simplest feature learning approach that was based on applying the K-means clustering algorithm after a whitening of the data. The goal of this work is to shed light on the success of K-means with whitening for the task of unsupervised feature learning. Our main result is a close connection between K-means and ICA (Independent Component Analysis). Specifically, we show that K-means and similar clustering algorithms can be used to recover the ICA mixing matrix or its inverse, the ICA filters. It is well known that the independent components found by ICA form useful features for classification (Le et al., 2012; 2011; 2010), hence the connection between K-mean and ICA explains the empirical success of K-means as a feature learner. Moreover, our analysis underscores the significance of the whitening operation, as was also observed in the experiments reported in (Coates et al., 2011). Finally, our analysis leads to a better initialization of K-means for the task of feature learning.