We introduce a weakly supervised method for representation learning based on aligning temporal sequences (e.g., videos) of the same process (e.g., human action). The main idea is to use the global temporal ordering of latent correspondences across sequence pairs as a supervisory signal. In particular, we propose a loss based on scoring the optimal sequence alignment to train an embedding network. Our loss is based on a novel probabilistic path finding view of dynamic time warping (DTW) that contains the following three key features: (i) the local path routing decisions are contrastive and differentiable, (ii) pairwise distances are cast as probabilities that are contrastive as well, and (iii) our formulation naturally admits a global cycle consistency loss that verifies correspondences. For evaluation, we consider the tasks of fine-grained action classification, few shot learning, and video synchronization. We report significant performance increases over previous methods. In addition, we report two applications of our temporal alignment framework, namely 3D pose reconstruction and fine-grained audio/visual retrieval.
We present a new method and a large-scale database to detect audio-video synchronization(A/V sync) errors in tennis videos. A deep network is trained to detect the visual signature of the tennis ball being hit by the racquet in the video stream. Another deep network is trained to detect the auditory signature of the same event in the audio stream. During evaluation, the audio stream is searched by the audio network for the audio event of the ball being hit. If the event is found in audio, the neighboring interval in video is searched for the corresponding visual signature. If the event is not found in the video stream but is found in the audio stream, A/V sync error is flagged. We developed a large-scaled database of 504,300 frames from 6 hours of videos of tennis events, simulated A/V sync errors, and found our method achieves high accuracy on the task.
Speaker diarization, which is to find the speech segments of specific speakers, has been widely used in human-centered applications such as video conferences or human-computer interaction systems. In this paper, we propose a self-supervised audio-video synchronization learning method to address the problem of speaker diarization without massive labeling effort. We improve the previous approaches by introducing two new loss functions: the dynamic triplet loss and the multinomial loss. We test them on a real-world human-computer interaction system and the results show our best model yields a remarkable gain of +8%F1-scoresas well as diarization error rate reduction. Finally, we introduce a new large scale audio-video corpus designed to fill the vacancy of audio-video datasets in Chinese.
Every minute, hundreds of hours of video are uploaded to social media sites and the Internet from around the world. This material creates a visual record of the experiences of a significant percentage of humanity and can help illuminate how we live in the present moment. When properly analyzed, this video can also help analysts to reconstruct events of interest, including war crimes, human rights violations, and terrorist acts. Machine learning and computer vision can play a crucial role in this process. In this technical report, we describe the Video Event Reconstruction and Analysis (VERA) system. This new tool brings together a variety of capabilities we have developed over the past few years (including video synchronization and geolocation to order unstructured videos lacking metadata over time and space, and sound recognition algorithms) to enable the reconstruction and analysis of events captured on video. Among other uses, VERA enables the localization of a shooter from just a few videos that include the sound of gunshots. To demonstrate the efficacy of this suite of tools, we present the results of estimating the shooter's location of the Las Vegas Shooting in 2017 and show that VERA accurately predicts the shooter's location using only the first few gunshots. We then point out future directions that can help improve the system and further reduce unnecessary human labor in the process. All of the components of VERA run through a web interface that enables human-in-the-loop verification to ensure accurate estimations. All relevant source code, including the web interface and machine learning models, is freely available on Github. We hope that researchers and software developers will be inspired to improve and expand this system moving forward to better meet the needs of human rights and public safety.
A simple method for synchronization of video streams with a precision better than one millisecond is proposed. The method is applicable to any number of rolling shutter cameras and when a few photographic flashes or other abrupt lighting changes are present in the video. The approach exploits the rolling shutter sensor property that every sensor row starts its exposure with a small delay after the onset of the previous row. The cameras may have different frame rates and resolutions, and need not have overlapping fields of view. The method was validated on five minutes of four streams from an ice hockey match. The found transformation maps events visible in all cameras to a reference time with a standard deviation of the temporal error in the range of 0.3 to 0.5 milliseconds. The quality of the synchronization is demonstrated on temporally and spatially overlapping images of a fast moving puck observed in two cameras.
This paper proposes a new strategy for learning powerful cross-modal embeddings for audio-to-video synchronization. Here, we set up the problem as one of cross-modal retrieval, where the objective is to find the most relevant audio segment given a short video clip. The method builds on the recent advances in learning representations from cross-modal self-supervision. The main contributions of this paper are as follows: (1) we propose a new learning strategy where the embeddings are learnt via a multi-way matching problem, as opposed to a binary classification (matching or non-matching) problem as proposed by recent papers; (2) we demonstrate that performance of this method far exceeds the existing baselines on the synchronization task; (3) we use the learnt embeddings for visual speech recognition in self-supervision, and show that the performance matches the representations learnt end-to-end in a fully-supervised manner.
Aligning video sequences is a fundamental yet still unsolved component for a broad range of applications in computer graphics and vision. Most classical image processing methods cannot be directly applied to related video problems due to the high amount of underlying data and their limit to small changes in appearance. We present a scalable and robust method for computing a non-linear temporal video alignment. The approach autonomously manages its training data for learning a meaningful representation in an iterative procedure each time increasing its own knowledge. It leverages on the nature of the videos themselves to remove the need for manually created labels. While previous alignment methods similarly consider weather conditions, season and illumination, our approach is able to align videos from data recorded months apart.