We present VideoPoet, a language model capable of synthesizing high-quality video from a large variety of conditioning signals. VideoPoet employs a decoder-only transformer architecture that processes multimodal inputs – including images, videos, text, and audio. The training protocol follows that of Large Language Models (LLMs), consisting of two stages: pretraining and task-specific adaptation. During pretraining, VideoPoet incorporates a mixture of multimodal generative objectives within an autoregressive Transformer framework. The pretrained LLM serves as a foundation that can be adapted for a range of video generation tasks. We present empirical results demonstrating the model’s state-of-the-art capabilities in zero-shot video generation, specifically highlighting the ability to generate high-fidelity motions. Project page: http://sites.research.google/videopoet/

We present data-driven techniques to augment Bag of Words (BoW) models, which allow for more robust modeling and recognition of complex long-term activities, especially when the structure and topology of the activities are not known a priori. Our approach specifically addresses the limitations of standard BoW approaches, which fail to represent the underlying temporal and causal information that is inherent in activity streams. In addition, we also propose the use of randomly sampled regular expressions to discover and encode patterns in activities. We demonstrate the effectiveness of our approach in experimental evaluations where we successfully recognize activities and detect anomalies in four complex datasets.

Activity recognition in the home has been long recognized as the foundation for many desirable applications in fields such as home automation, sustainability, and healthcare. However, building a practical home activity monitoring system remains a challenge. Striking a balance between cost, privacy, ease of installation and scalability continues to be an elusive goal. In this paper, we explore infrastructure-mediated sensing combined with a vector space model learning approach as the basis of an activity recognition system for the home. We examine the performance of our single-sensor water-based system in recognizing eleven high-level activities in the kitchen and bathroom, such as cooking and shaving. Results from two studies show that our system can estimate activities with overall accuracy of 82.69% for one individual and 70.11% for a group of 23 participants. As far as we know, our work is the first to employ infrastructure-mediated sensing for inferring high-level human activities in a home setting.

We present a novel approach that allows anyone to quickly teach their smartphone how to understand the visual world around them. We achieve this visual scene understanding by leveraging a camera-phone's inertial sensors to lead to both a faster and more accurate automatic labeling of the regions of an image into semantic classes (e.g. sky, tree, building). We focus on letting a user train our system from scratch while out in the real world by annotating image regions in situ as training images are captured on a mobile device, making it possible to recognize new environments and new semantic classes on the fly. We show that our approach outperforms existing methods, while at the same time performing data collection, annotation, feature extraction, and image segment classification all on the same mobile device.

Modern structure from motion techniques are capable of building city-scale 3D reconstructions from large image collections, but have mostly ignored the problem of large-scale structural changes over time. We present a general framework for estimating temporal variables in structure from motion problems, including an unknown date for each camera and an unknown time interval for each structural element. Given a collection of images with mostly unknown or uncertain dates, we use this framework to automatically recover the dates of all images by reasoning probabilistically about the visibility and existence of objects in the scene. We present results on a collection of over 100 historical images of a city taken over decades of time.

We present a novel method for automatically geo-tagging photographs of man-made environments via detection and matching of repeated patterns. Highly repetitive environments introduce numerous correspondence ambiguities and are problematic for traditional wide-baseline matching methods. Our method exploits the highly repetitive nature of urban environments, detecting multiple perspectively distorted periodic 2D patterns in an image and matching them to a 3D database of textured facades by reasoning about the underlying canonical forms of each pattern. Multiple 2D-to-3D pattern correspondences enable robust recovery of camera orientation and location. We demonstrate the success of this method in a large urban environment.

We look at the problem of location recognition in a large image dataset using a vocabulary tree. This entails finding the location of a query image in a large dataset containing 3times104streetside images of a city. We investigate how the traditional invariant feature matching approach falls down as the size of the database grows. In particular we show that by carefully selecting the vocabulary using the most informative features, retrieval performance is significantly improved, allowing us to increase the number of database images by a factor of 10. We also introduce a generalization of the traditional vocabulary tree search algorithm which improves performance by effectively increasing the branching factor of a fixed vocabulary tree.

In this paper, we describe a technique to temporally sort a collection of photos that span many years. By reasoning about persistence of visible structures, we show how this sorting task can be formulated as a constraint satisfaction problem (CSP). Casting this problem as a CSP allows us to efficiently find a suitable ordering of the images despite the large size of the solution space (factorial in the number of images) and the presence of occlusions. We present experimental results for photographs of a city acquired over a one hundred year period.