Bryan Russell

Hang Gao, Ruilong Li, Shubham Tulsiani et al.

We study the recent progress on dynamic view synthesis (DVS) from monocular video. Though existing approaches have demonstrated impressive results, we show a discrepancy between the practical capture process and the existing experimental protocols, which effectively leaks in multi-view signals during training. We define effective multi-view factors (EMFs) to quantify the amount of multi-view signal present in the input capture sequence based on the relative camera-scene motion. We introduce two new metrics: co-visibility masked image metrics and correspondence accuracy, which overcome the issue in existing protocols. We also propose a new iPhone dataset that includes more diverse real-life deformation sequences. Using our proposed experimental protocol, we show that the state-of-the-art approaches observe a 1-2 dB drop in masked PSNR in the absence of multi-view cues and 4-5 dB drop when modeling complex motion. Code and data can be found at https://hangg7.com/dycheck.

Zhongzheng Ren, Aseem Agarwala, Bryan Russell et al.

We introduce an approach for selecting objects in neural volumetric 3D representations, such as multi-plane images (MPI) and neural radiance fields (NeRF). Our approach takes a set of foreground and background 2D user scribbles in one view and automatically estimates a 3D segmentation of the desired object, which can be rendered into novel views. To achieve this result, we propose a novel voxel feature embedding that incorporates the neural volumetric 3D representation and multi-view image features from all input views. To evaluate our approach, we introduce a new dataset of human-provided segmentation masks for depicted objects in real-world multi-view scene captures. We show that our approach out-performs strong baselines, including 2D segmentation and 3D segmentation approaches adapted to our task.

Reuben Tan, Arijit Ray, Andrea Burns et al.

We propose a self-supervised approach for learning to perform audio source separation in videos based on natural language queries, using only unlabeled video and audio pairs as training data. A key challenge in this task is learning to associate the linguistic description of a sound-emitting object to its visual features and the corresponding components of the audio waveform, all without access to annotations during training. To overcome this challenge, we adapt off-the-shelf vision-language foundation models to provide pseudo-target supervision via two novel loss functions and encourage a stronger alignment between the audio, visual and natural language modalities. During inference, our approach can separate sounds given text, video and audio input, or given text and audio input alone. We demonstrate the effectiveness of our self-supervised approach on three audio-visual separation datasets, including MUSIC, SOLOS and AudioSet, where we outperform state-of-the-art strongly supervised approaches despite not using object detectors or text labels during training.

Daniel McKee, Justin Salamon, Josef Sivic et al.

We propose a method to recommend music for an input video while allowing a user to guide music selection with free-form natural language. A key challenge of this problem setting is that existing music video datasets provide the needed (video, music) training pairs, but lack text descriptions of the music. This work addresses this challenge with the following three contributions. First, we propose a text-synthesis approach that relies on an analogy-based prompting procedure to generate natural language music descriptions from a large-scale language model (BLOOM-176B) given pre-trained music tagger outputs and a small number of human text descriptions. Second, we use these synthesized music descriptions to train a new trimodal model, which fuses text and video input representations to query music samples. For training, we introduce a text dropout regularization mechanism which we show is critical to model performance. Our model design allows for the retrieved music audio to agree with the two input modalities by matching visual style depicted in the video and musical genre, mood, or instrumentation described in the natural language query. Third, to evaluate our approach, we collect a testing dataset for our problem by annotating a subset of 4k clips from the YT8M-MusicVideo dataset with natural language music descriptions which we make publicly available. We show that our approach can match or exceed the performance of prior methods on video-to-music retrieval while significantly improving retrieval accuracy when using text guidance.

Georgy Ponimatkin, Yann Labbé, Bryan Russell et al.

We introduce FocalPose, a neural render-and-compare method for jointly estimating the camera-object 6D pose and camera focal length given a single RGB input image depicting a known object. The contributions of this work are twofold. First, we derive a focal length update rule that extends an existing state-of-the-art render-and-compare 6D pose estimator to address the joint estimation task. Second, we investigate several different loss functions for jointly estimating the object pose and focal length. We find that a combination of direct focal length regression with a reprojection loss disentangling the contribution of translation, rotation, and focal length leads to improved results. We show results on three challenging benchmark datasets that depict known 3D models in uncontrolled settings. We demonstrate that our focal length and 6D pose estimates have lower error than the existing state-of-the-art methods.

Chun-Hsiao Yeh, Bryan Russell, Josef Sivic et al.

Large-scale vision-language models (VLM) have shown impressive results for language-guided search applications. While these models allow category-level queries, they currently struggle with personalized searches for moments in a video where a specific object instance such as ``My dog Biscuit'' appears. We present the following three contributions to address this problem. First, we describe a method to meta-personalize a pre-trained VLM, i.e., learning how to learn to personalize a VLM at test time to search in video. Our method extends the VLM's token vocabulary by learning novel word embeddings specific to each instance. To capture only instance-specific features, we represent each instance embedding as a combination of shared and learned global category features. Second, we propose to learn such personalization without explicit human supervision. Our approach automatically identifies moments of named visual instances in video using transcripts and vision-language similarity in the VLM's embedding space. Finally, we introduce This-Is-My, a personal video instance retrieval benchmark. We evaluate our approach on This-Is-My and DeepFashion2 and show that we obtain a 15% relative improvement over the state of the art on the latter dataset.

Yuexi Du, Ziyang Chen, Justin Salamon et al.

The sound effects that designers add to videos are designed to convey a particular artistic effect and, thus, may be quite different from a scene's true sound. Inspired by the challenges of creating a soundtrack for a video that differs from its true sound, but that nonetheless matches the actions occurring on screen, we propose the problem of conditional Foley. We present the following contributions to address this problem. First, we propose a pretext task for training our model to predict sound for an input video clip using a conditional audio-visual clip sampled from another time within the same source video. Second, we propose a model for generating a soundtrack for a silent input video, given a user-supplied example that specifies what the video should "sound like". We show through human studies and automated evaluation metrics that our model successfully generates sound from video, while varying its output according to the content of a supplied example. Project site: https://xypb.github.io/CondFoleyGen/

Didac Suris, Carl Vondrick, Bryan Russell et al.

We present an approach for recommending a music track for a given video, and vice versa, based on both their temporal alignment and their correspondence at an artistic level. We propose a self-supervised approach that learns this correspondence directly from data, without any need of human annotations. In order to capture the high-level concepts that are required to solve the task, we propose modeling the long-term temporal context of both the video and the music signals, using Transformer networks for each modality. Experiments show that this approach strongly outperforms alternatives that do not exploit the temporal context. The combination of our contributions improve retrieval accuracy up to 10x over prior state of the art. This strong improvement allows us to introduce a wide range of analyses and applications. For instance, we can condition music retrieval based on visually defined attributes.

Martin Cífka, Georgy Ponimatkin, Yann Labbé et al.

We introduce FocalPose++, a neural render-and-compare method for jointly estimating the camera-object 6D pose and camera focal length given a single RGB input image depicting a known object. The contributions of this work are threefold. First, we derive a focal length update rule that extends an existing state-of-the-art render-and-compare 6D pose estimator to address the joint estimation task. Second, we investigate several different loss functions for jointly estimating the object pose and focal length. We find that a combination of direct focal length regression with a reprojection loss disentangling the contribution of translation, rotation, and focal length leads to improved results. Third, we explore the effect of different synthetic training data on the performance of our method. Specifically, we investigate different distributions used for sampling object's 6D pose and camera's focal length when rendering the synthetic images, and show that parametric distribution fitted on real training data works the best. We show results on three challenging benchmark datasets that depict known 3D models in uncontrolled settings. We demonstrate that our focal length and 6D pose estimates have lower error than the existing state-of-the-art methods.

Lisa Dunlap, Joseph E. Gonzalez, Trevor Darrell et al.

In this paper, we investigate when and how visual representations learned by two different generative models diverge. Given two text-to-image models, our goal is to discover visual attributes that appear in images generated by one model but not the other, along with the types of prompts that trigger these attribute differences. For example, "flames" might appear in one model's outputs when given prompts expressing strong emotions, while the other model does not produce this attribute given the same prompts. We introduce CompCon (Comparing Concepts), an evolutionary search algorithm that discovers visual attributes more prevalent in one model's output than the other, and uncovers the prompt concepts linked to these visual differences. To evaluate CompCon's ability to find diverging representations, we create an automated data generation pipeline to produce ID2, a dataset of 60 input-dependent differences, and compare our approach to several LLM- and VLM-powered baselines. Finally, we use CompCon to compare popular text-to-image models, finding divergent representations such as how PixArt depicts prompts mentioning loneliness with wet streets and Stable Diffusion 3.5 depicts African American people in media professions. Code at: https://github.com/adobe-research/CompCon

Reuben Tan, Ximeng Sun, Ping Hu et al.

Long video question answering is a challenging task that involves recognizing short-term activities and reasoning about their fine-grained relationships. State-of-the-art video Large Language Models (vLLMs) hold promise as a viable solution due to their demonstrated emergent capabilities on new tasks. However, despite being trained on millions of short seconds-long videos, vLLMs are unable to understand minutes-long videos and accurately answer questions about them. To address this limitation, we propose a lightweight and self-supervised approach, Key frame-conditioned long video-LLM (Koala), that introduces learnable spatiotemporal queries to adapt pretrained vLLMs for generalizing to longer videos. Our approach introduces two new tokenizers that condition on visual tokens computed from sparse video key frames for understanding short and long video moments. We train our proposed approach on HowTo100M and demonstrate its effectiveness on zero-shot long video understanding benchmarks, where it outperforms state-of-the-art large models by 3 - 6% in absolute accuracy across all tasks. Surprisingly, we also empirically show that our approach not only helps a pretrained vLLM to understand long videos but also improves its accuracy on short-term action recognition.

Ziyang Chen, Prem Seetharaman, Bryan Russell et al.

Generating sound effects for videos often requires creating artistic sound effects that diverge significantly from real-life sources and flexible control in the sound design. To address this problem, we introduce MultiFoley, a model designed for video-guided sound generation that supports multimodal conditioning through text, audio, and video. Given a silent video and a text prompt, MultiFoley allows users to create clean sounds (e.g., skateboard wheels spinning without wind noise) or more whimsical sounds (e.g., making a lion's roar sound like a cat's meow). MultiFoley also allows users to choose reference audio from sound effects (SFX) libraries or partial videos for conditioning. A key novelty of our model lies in its joint training on both internet video datasets with low-quality audio and professional SFX recordings, enabling high-quality, full-bandwidth (48kHz) audio generation. Through automated evaluations and human studies, we demonstrate that MultiFoley successfully generates synchronized high-quality sounds across varied conditional inputs and outperforms existing methods. Please see our project page for video results: https://ificl.github.io/MultiFoley/

Joanna Materzynska, Josef Sivic, Eli Shechtman et al.

We introduce an approach for augmenting text-to-video generation models with customized motions, extending their capabilities beyond the motions depicted in the original training data. By leveraging a few video samples demonstrating specific movements as input, our method learns and generalizes the input motion patterns for diverse, text-specified scenarios. Our contributions are threefold. First, to achieve our results, we finetune an existing text-to-video model to learn a novel mapping between the depicted motion in the input examples to a new unique token. To avoid overfitting to the new custom motion, we introduce an approach for regularization over videos. Second, by leveraging the motion priors in a pretrained model, our method can produce novel videos featuring multiple people doing the custom motion, and can invoke the motion in combination with other motions. Furthermore, our approach extends to the multimodal customization of motion and appearance of individualized subjects, enabling the generation of videos featuring unique characters and distinct motions. Third, to validate our method, we introduce an approach for quantitatively evaluating the learned custom motion and perform a systematic ablation study. We show that our method significantly outperforms prior appearance-based customization approaches when extended to the motion customization task.

Mattia Soldan, Fabian Caba Heilbron, Bernard Ghanem et al.

Several video understanding tasks, such as natural language temporal video grounding, temporal activity localization, and audio description generation, require "temporally dense" reasoning over frames sampled at high temporal resolution. However, computing frame-level features for these tasks is computationally expensive given the temporal resolution requirements. In this paper, we make three contributions to reduce the cost of computing features for temporally dense tasks. First, we introduce a vision transformer (ViT) architecture, dubbed ResidualViT, that leverages the large temporal redundancy in videos to efficiently compute temporally dense frame-level features. Our architecture incorporates (i) learnable residual connections that ensure temporal consistency across consecutive frames and (ii) a token reduction module that enhances processing speed by selectively discarding temporally redundant information while reusing weights of a pretrained foundation model. Second, we propose a lightweight distillation strategy to approximate the frame-level features of the original foundation model. Finally, we evaluate our approach across four tasks and five datasets, in both zero-shot and fully supervised settings, demonstrating significant reductions in computational cost (up to 60%) and improvements in inference speed (up to 2.5x faster), all while closely approximating the accuracy of the original foundation model.

Marcelo Sandoval-Castaneda, Bryan Russell, Josef Sivic et al.

Automated tools for video editing and assembly have applications ranging from filmmaking and advertisement to content creation for social media. Previous video editing work has mainly focused on either retrieval or user interfaces, leaving actual editing to the user. In contrast, we propose to automate the core task of video editing, formulating it as sequential decision making process. Ours is a multi-agent approach. We design an Editor agent and a Critic agent. The Editor takes as input a collection of video clips together with natural language instructions and uses tools commonly found in video editing software to produce an edited sequence. On the other hand, the Critic gives natural language feedback to the editor based on the produced sequence or renders it if it is satisfactory. We introduce a learning-based approach for enabling effective communication across specialized agents to address the language-driven video editing task. Finally, we explore an LLM-as-a-judge metric for evaluating the quality of video editing system and compare it with general human preference. We evaluate our system's output video sequences qualitatively and quantitatively through a user study and find that our system vastly outperforms existing approaches in terms of coverage, time constraint satisfaction, and human preference.

Fiona Ryan, Josef Sivic, Fabian Caba Heilbron et al. · gatech

Personalized vision-language retrieval seeks to recognize new concepts (e.g. "my dog Fido") from only a few examples. This task is challenging because it requires not only learning a new concept from a few images, but also integrating the personal and general knowledge together to recognize the concept in different contexts. In this paper, we show how to effectively adapt the internal representation of a vision-language dual encoder model for personalized vision-language retrieval. We find that regularized low-rank adaption of a small set of parameters in the language encoder's final layer serves as a highly effective alternative to textual inversion for recognizing the personal concept while preserving general knowledge. Additionally, we explore strategies for combining parameters of multiple learned personal concepts, finding that parameter addition is effective. To evaluate how well general knowledge is preserved in a finetuned representation, we introduce a metric that measures image retrieval accuracy based on captions generated by a vision language model (VLM). Our approach achieves state-of-the-art accuracy on two benchmarks for personalized image retrieval with natural language queries - DeepFashion2 and ConCon-Chi - outperforming the prior art by 4%-22% on personal retrievals.

Alejandro Pardo, Jui-Hsien Wang, Bernard Ghanem et al.

The objective of this work is to manipulate visual timelines (e.g. a video) through natural language instructions, making complex timeline editing tasks accessible to non-expert or potentially even disabled users. We call this task Instructed visual assembly. This task is challenging as it requires (i) identifying relevant visual content in the input timeline as well as retrieving relevant visual content in a given input (video) collection, (ii) understanding the input natural language instruction, and (iii) performing the desired edits of the input visual timeline to produce an output timeline. To address these challenges, we propose the Timeline Assembler, a generative model trained to perform instructed visual assembly tasks. The contributions of this work are three-fold. First, we develop a large multimodal language model, which is designed to process visual content, compactly represent timelines and accurately interpret timeline editing instructions. Second, we introduce a novel method for automatically generating datasets for visual assembly tasks, enabling efficient training of our model without the need for human-labeled data. Third, we validate our approach by creating two novel datasets for image and video assembly, demonstrating that the Timeline Assembler substantially outperforms established baseline models, including the recent GPT-4o, in accurately executing complex assembly instructions across various real-world inspired scenarios.

Jiacheng Cheng, Hijung Valentina Shin, Nuno Vasconcelos et al.

In the recent years, the dual-encoder vision-language models (\eg CLIP) have achieved remarkable text-to-image retrieval performance. However, we discover that these models usually results in very different retrievals for a pair of paraphrased queries. Such behavior might render the retrieval system less predictable and lead to user frustration. In this work, we consider the task of paraphrased text-to-image retrieval where a model aims to return similar results given a pair of paraphrased queries. To start with, we collect a dataset of paraphrased image descriptions to facilitate quantitative evaluation for this task. We then hypothesize that the undesired behavior of existing dual-encoder model is due to their text towers which are trained on image-sentence pairs and lack the ability to capture the semantic similarity between paraphrased queries. To improve on this, we investigate multiple strategies for training a dual-encoder model starting from a language model pretrained on a large text corpus. Compared to public dual-encoder models such as CLIP and OpenCLIP, the model trained with our best adaptation strategy achieves a significantly higher ranking similarity for paraphrased queries while maintaining similar zero-shot classification and retrieval accuracy.

Alex Andonian, Taesung Park, Bryan Russell et al.

Training supervised image synthesis models requires a critic to compare two images: the ground truth to the result. Yet, this basic functionality remains an open problem. A popular line of approaches uses the L1 (mean absolute error) loss, either in the pixel or the feature space of pretrained deep networks. However, we observe that these losses tend to produce overly blurry and grey images, and other techniques such as GANs need to be employed to fight these artifacts. In this work, we introduce an information theory based approach to measuring similarity between two images. We argue that a good reconstruction should have high mutual information with the ground truth. This view enables learning a lightweight critic to "calibrate" a feature space in a contrastive manner, such that reconstructions of corresponding spatial patches are brought together, while other patches are repulsed. We show that our formulation immediately boosts the perceptual realism of output images when used as a drop-in replacement for the L1 loss, with or without an additional GAN loss.

Reuben Tan, Bryan A. Plummer, Kate Saenko et al.

We introduce the task of spatially localizing narrated interactions in videos. Key to our approach is the ability to learn to spatially localize interactions with self-supervision on a large corpus of videos with accompanying transcribed narrations. To achieve this goal, we propose a multilayer cross-modal attention network that enables effective optimization of a contrastive loss during training. We introduce a divided strategy that alternates between computing inter- and intra-modal attention across the visual and natural language modalities, which allows effective training via directly contrasting the two modalities' representations. We demonstrate the effectiveness of our approach by self-training on the HowTo100M instructional video dataset and evaluating on a newly collected dataset of localized described interactions in the YouCook2 dataset. We show that our approach outperforms alternative baselines, including shallow co-attention and full cross-modal attention. We also apply our approach to grounding phrases in images with weak supervision on Flickr30K and show that stacking multiple attention layers is effective and, when combined with a word-to-region loss, achieves state of the art on recall-at-one and pointing hand accuracies.

Shuang Li, Yilun Du, Antonio Torralba et al.

We introduce the task of weakly supervised learning for detecting human and object interactions in videos. Our task poses unique challenges as a system does not know what types of human-object interactions are present in a video or the actual spatiotemporal location of the human and the object. To address these challenges, we introduce a contrastive weakly supervised training loss that aims to jointly associate spatiotemporal regions in a video with an action and object vocabulary and encourage temporal continuity of the visual appearance of moving objects as a form of self-supervision. To train our model, we introduce a dataset comprising over 6.5k videos with human-object interaction annotations that have been semi-automatically curated from sentence captions associated with the videos. We demonstrate improved performance over weakly supervised baselines adapted to our task on our video dataset.

Steven Liu, Xiuming Zhang, Zhoutong Zhang et al.

A neural radiance field (NeRF) is a scene model supporting high-quality view synthesis, optimized per scene. In this paper, we explore enabling user editing of a category-level NeRF - also known as a conditional radiance field - trained on a shape category. Specifically, we introduce a method for propagating coarse 2D user scribbles to the 3D space, to modify the color or shape of a local region. First, we propose a conditional radiance field that incorporates new modular network components, including a shape branch that is shared across object instances. Observing multiple instances of the same category, our model learns underlying part semantics without any supervision, thereby allowing the propagation of coarse 2D user scribbles to the entire 3D region (e.g., chair seat). Next, we propose a hybrid network update strategy that targets specific network components, which balances efficiency and accuracy. During user interaction, we formulate an optimization problem that both satisfies the user's constraints and preserves the original object structure. We demonstrate our approach on various editing tasks over three shape datasets and show that it outperforms prior neural editing approaches. Finally, we edit the appearance and shape of a real photograph and show that the edit propagates to extrapolated novel views.

Davis Rempe, Leonidas J. Guibas, Aaron Hertzmann et al.

Existing deep models predict 2D and 3D kinematic poses from video that are approximately accurate, but contain visible errors that violate physical constraints, such as feet penetrating the ground and bodies leaning at extreme angles. In this paper, we present a physics-based method for inferring 3D human motion from video sequences that takes initial 2D and 3D pose estimates as input. We first estimate ground contact timings with a novel prediction network which is trained without hand-labeled data. A physics-based trajectory optimization then solves for a physically-plausible motion, based on the inputs. We show this process produces motions that are significantly more realistic than those from purely kinematic methods, substantially improving quantitative measures of both kinematic and dynamic plausibility. We demonstrate our method on character animation and pose estimation tasks on dynamic motions of dancing and sports with complex contact patterns.

Karren Yang, Bryan Russell, Justin Salamon

Self-supervised audio-visual learning aims to capture useful representations of video by leveraging correspondences between visual and audio inputs. Existing approaches have focused primarily on matching semantic information between the sensory streams. We propose a novel self-supervised task to leverage an orthogonal principle: matching spatial information in the audio stream to the positions of sound sources in the visual stream. Our approach is simple yet effective. We train a model to determine whether the left and right audio channels have been flipped, forcing it to reason about spatial localization across the visual and audio streams. To train and evaluate our method, we introduce a large-scale video dataset, YouTube-ASMR-300K, with spatial audio comprising over 900 hours of footage. We demonstrate that understanding spatial correspondence enables models to perform better on three audio-visual tasks, achieving quantitative gains over supervised and self-supervised baselines that do not leverage spatial audio cues. We also show how to extend our self-supervised approach to 360 degree videos with ambisonic audio.

Christian Zimmermann, Duygu Ceylan, Jimei Yang et al.

Estimating 3D hand pose from single RGB images is a highly ambiguous problem that relies on an unbiased training dataset. In this paper, we analyze cross-dataset generalization when training on existing datasets. We find that approaches perform well on the datasets they are trained on, but do not generalize to other datasets or in-the-wild scenarios. As a consequence, we introduce the first large-scale, multi-view hand dataset that is accompanied by both 3D hand pose and shape annotations. For annotating this real-world dataset, we propose an iterative, semi-automated `human-in-the-loop' approach, which includes hand fitting optimization to infer both the 3D pose and shape for each sample. We show that methods trained on our dataset consistently perform well when tested on other datasets. Moreover, the dataset allows us to train a network that predicts the full articulated hand shape from a single RGB image. The evaluation set can serve as a benchmark for articulated hand shape estimation.

Carlo Innamorati, Bryan Russell, Danny M. Kaufman et al.

We introduce a method to generate videos of dynamic virtual objects plausibly interacting via collisions with a still image's environment. Given a starting trajectory, physically simulated with the estimated geometry of a single, static input image, we learn to 'correct' this trajectory to a visually plausible one via a neural network. The neural network can then be seen as learning to 'correct' traditional simulation output, generated with incomplete and imprecise world information, to obtain context-specific, visually plausible re-simulated output, a process we call neural re-simulation. We train our system on a set of 50k synthetic scenes where a virtual moving object (ball) has been physically simulated. We demonstrate our approach on both our synthetic dataset and a collection of real-life images depicting everyday scenes, obtaining consistent improvement over baseline alternatives throughout.

Victor Escorcia, Mattia Soldan, Josef Sivic et al.

We introduce the task of retrieving relevant video moments from a large corpus of untrimmed, unsegmented videos given a natural language query. Our task poses unique challenges as a system must efficiently identify both the relevant videos and localize the relevant moments in the videos. To address these challenges, we propose SpatioTemporal Alignment with Language (STAL), a model that represents a video moment as a set of regions within a series of short video clips and aligns a natural language query to the moment's regions. Our alignment cost compares variable-length language and video features using symmetric squared Chamfer distance, which allows for efficient indexing and retrieval of the video moments. Moreover, aligning language features to regions within a video moment allows for finer alignment compared to methods that extract only an aggregate feature from the entire video moment. We evaluate our approach on two recently proposed datasets for temporal localization of moments in video with natural language (DiDeMo and Charades-STA) extended to our video corpus moment retrieval setting. We show that our STAL re-ranking model outperforms the recently proposed Moment Context Network on all criteria across all datasets on our proposed task, obtaining relative gains of 37% - 118% for average recall and up to 30% for median rank. Moreover, our approach achieves more than 130x faster retrieval and 8x smaller index size with a 1M video corpus in an approximate setting.

Senthil Purushwalkam, Abhinav Gupta, Danny M. Kaufman et al.

We introduce an approach to model surface properties governing bounces in everyday scenes. Our model learns end-to-end, starting from sensor inputs, to predict post-bounce trajectories and infer two underlying physical properties that govern bouncing - restitution and effective collision normals. Our model, Bounce and Learn, comprises two modules -- a Physics Inference Module (PIM) and a Visual Inference Module (VIM). VIM learns to infer physical parameters for locations in a scene given a single still image, while PIM learns to model physical interactions for the prediction task given physical parameters and observed pre-collision 3D trajectories. To achieve our results, we introduce the Bounce Dataset comprising 5K RGB-D videos of bouncing trajectories of a foam ball to probe surfaces of varying shapes and materials in everyday scenes including homes and offices. Our proposed model learns from our collected dataset of real-world bounces and is bootstrapped with additional information from simple physics simulations. We show on our newly collected dataset that our model out-performs baselines, including trajectory fitting with Newtonian physics, in predicting post-bounce trajectories and inferring physical properties of a scene.

Bernd Huber, Hijung Valentina Shin, Bryan Russell et al.

In video production, inserting B-roll is a widely used technique to enrich the story and make a video more engaging. However, determining the right content and positions of B-roll and actually inserting it within the main footage can be challenging, and novice producers often struggle to get both timing and content right. We present B-Script, a system that supports B-roll video editing via interactive transcripts. B-Script has a built-in recommendation system trained on expert-annotated data, recommending users B-roll position and content. To evaluate the system, we conducted a within-subject user study with 110 participants, and compared three interface variations: a timeline-based editor, a transcript-based editor, and a transcript-based editor with recommendations. Users found it easier and were faster to insert B-roll using the transcript-based interface, and they created more engaging videos when recommendations were provided.

Lisa Anne Hendricks, Oliver Wang, Eli Shechtman et al.

Localizing moments in a longer video via natural language queries is a new, challenging task at the intersection of language and video understanding. Though moment localization with natural language is similar to other language and vision tasks like natural language object retrieval in images, moment localization offers an interesting opportunity to model temporal dependencies and reasoning in text. We propose a new model that explicitly reasons about different temporal segments in a video, and shows that temporal context is important for localizing phrases which include temporal language. To benchmark whether our model, and other recent video localization models, can effectively reason about temporal language, we collect the novel TEMPOral reasoning in video and language (TEMPO) dataset. Our dataset consists of two parts: a dataset with real videos and template sentences (TEMPO - Template Language) which allows for controlled studies on temporal language, and a human language dataset which consists of temporal sentences annotated by humans (TEMPO - Human Language).

Gül Varol, Duygu Ceylan, Bryan Russell et al.

Human shape estimation is an important task for video editing, animation and fashion industry. Predicting 3D human body shape from natural images, however, is highly challenging due to factors such as variation in human bodies, clothing and viewpoint. Prior methods addressing this problem typically attempt to fit parametric body models with certain priors on pose and shape. In this work we argue for an alternative representation and propose BodyNet, a neural network for direct inference of volumetric body shape from a single image. BodyNet is an end-to-end trainable network that benefits from (i) a volumetric 3D loss, (ii) a multi-view re-projection loss, and (iii) intermediate supervision of 2D pose, 2D body part segmentation, and 3D pose. Each of them results in performance improvement as demonstrated by our experiments. To evaluate the method, we fit the SMPL model to our network output and show state-of-the-art results on the SURREAL and Unite the People datasets, outperforming recent approaches. Besides achieving state-of-the-art performance, our method also enables volumetric body-part segmentation.

Zoya Bylinskii, Nam Wook Kim, Peter O'Donovan et al.

Knowing where people look and click on visual designs can provide clues about how the designs are perceived, and where the most important or relevant content lies. The most important content of a visual design can be used for effective summarization or to facilitate retrieval from a database. We present automated models that predict the relative importance of different elements in data visualizations and graphic designs. Our models are neural networks trained on human clicks and importance annotations on hundreds of designs. We collected a new dataset of crowdsourced importance, and analyzed the predictions of our models with respect to ground truth importance and human eye movements. We demonstrate how such predictions of importance can be used for automatic design retargeting and thumbnailing. User studies with hundreds of MTurk participants validate that, with limited post-processing, our importance-driven applications are on par with, or outperform, current state-of-the-art methods, including natural image saliency. We also provide a demonstration of how our importance predictions can be built into interactive design tools to offer immediate feedback during the design process.

Lisa Anne Hendricks, Oliver Wang, Eli Shechtman et al.

We consider retrieving a specific temporal segment, or moment, from a video given a natural language text description. Methods designed to retrieve whole video clips with natural language determine what occurs in a video but not when. To address this issue, we propose the Moment Context Network (MCN) which effectively localizes natural language queries in videos by integrating local and global video features over time. A key obstacle to training our MCN model is that current video datasets do not include pairs of localized video segments and referring expressions, or text descriptions which uniquely identify a corresponding moment. Therefore, we collect the Distinct Describable Moments (DiDeMo) dataset which consists of over 10,000 unedited, personal videos in diverse visual settings with pairs of localized video segments and referring expressions. We demonstrate that MCN outperforms several baseline methods and believe that our initial results together with the release of DiDeMo will inspire further research on localizing video moments with natural language.

Rohit Girdhar, Deva Ramanan, Abhinav Gupta et al.

In this work, we introduce a new video representation for action classification that aggregates local convolutional features across the entire spatio-temporal extent of the video. We do so by integrating state-of-the-art two-stream networks with learnable spatio-temporal feature aggregation. The resulting architecture is end-to-end trainable for whole-video classification. We investigate different strategies for pooling across space and time and combining signals from the different streams. We find that: (i) it is important to pool jointly across space and time, but (ii) appearance and motion streams are best aggregated into their own separate representations. Finally, we show that our representation outperforms the two-stream base architecture by a large margin (13% relative) as well as out-performs other baselines with comparable base architectures on HMDB51, UCF101, and Charades video classification benchmarks.

Aayush Bansal, Xinlei Chen, Bryan Russell et al.

We explore design principles for general pixel-level prediction problems, from low-level edge detection to mid-level surface normal estimation to high-level semantic segmentation. Convolutional predictors, such as the fully-convolutional network (FCN), have achieved remarkable success by exploiting the spatial redundancy of neighboring pixels through convolutional processing. Though computationally efficient, we point out that such approaches are not statistically efficient during learning precisely because spatial redundancy limits the information learned from neighboring pixels. We demonstrate that stratified sampling of pixels allows one to (1) add diversity during batch updates, speeding up learning; (2) explore complex nonlinear predictors, improving accuracy; and (3) efficiently train state-of-the-art models tabula rasa (i.e., "from scratch") for diverse pixel-labeling tasks. Our single architecture produces state-of-the-art results for semantic segmentation on PASCAL-Context dataset, surface normal estimation on NYUDv2 depth dataset, and edge detection on BSDS.

Aayush Bansal, Xinlei Chen, Bryan Russell et al.

We explore architectures for general pixel-level prediction problems, from low-level edge detection to mid-level surface normal estimation to high-level semantic segmentation. Convolutional predictors, such as the fully-convolutional network (FCN), have achieved remarkable success by exploiting the spatial redundancy of neighboring pixels through convolutional processing. Though computationally efficient, we point out that such approaches are not statistically efficient during learning precisely because spatial redundancy limits the information learned from neighboring pixels. We demonstrate that (1) stratified sampling allows us to add diversity during batch updates and (2) sampled multi-scale features allow us to explore more nonlinear predictors (multiple fully-connected layers followed by ReLU) that improve overall accuracy. Finally, our objective is to show how a architecture can get performance better than (or comparable to) the architectures designed for a particular task. Interestingly, our single architecture produces state-of-the-art results for semantic segmentation on PASCAL-Context, surface normal estimation on NYUDv2 dataset, and edge detection on BSDS without contextual post-processing.

Aayush Bansal, Bryan Russell, Abhinav Gupta

We introduce an approach that leverages surface normal predictions, along with appearance cues, to retrieve 3D models for objects depicted in 2D still images from a large CAD object library. Critical to the success of our approach is the ability to recover accurate surface normals for objects in the depicted scene. We introduce a skip-network model built on the pre-trained Oxford VGG convolutional neural network (CNN) for surface normal prediction. Our model achieves state-of-the-art accuracy on the NYUv2 RGB-D dataset for surface normal prediction, and recovers fine object detail compared to previous methods. Furthermore, we develop a two-stream network over the input image and predicted surface normals that jointly learns pose and style for CAD model retrieval. When using the predicted surface normals, our two-stream network matches prior work using surface normals computed from RGB-D images on the task of pose prediction, and achieves state of the art when using RGB-D input. Finally, our two-stream network allows us to retrieve CAD models that better match the style and pose of a depicted object compared with baseline approaches.

Francisco Massa, Bryan Russell, Mathieu Aubry

This paper presents an end-to-end convolutional neural network (CNN) for 2D-3D exemplar detection. We demonstrate that the ability to adapt the features of natural images to better align with those of CAD rendered views is critical to the success of our technique. We show that the adaptation can be learned by compositing rendered views of textured object models on natural images. Our approach can be naturally incorporated into a CNN detection pipeline and extends the accuracy and speed benefits from recent advances in deep learning to 2D-3D exemplar detection. We applied our method to two tasks: instance detection, where we evaluated on the IKEA dataset, and object category detection, where we out-perform Aubry et al. for "chair" detection on a subset of the Pascal VOC dataset.

Mathieu Aubry, Bryan Russell

We introduce an approach for analyzing the variation of features generated by convolutional neural networks (CNNs) with respect to scene factors that occur in natural images. Such factors may include object style, 3D viewpoint, color, and scene lighting configuration. Our approach analyzes CNN feature responses corresponding to different scene factors by controlling for them via rendering using a large database of 3D CAD models. The rendered images are presented to a trained CNN and responses for different layers are studied with respect to the input scene factors. We perform a decomposition of the responses based on knowledge of the input scene factors and analyze the resulting components. In particular, we quantify their relative importance in the CNN responses and visualize them using principal component analysis. We show qualitative and quantitative results of our study on three CNNs trained on large image datasets: AlexNet, Places, and Oxford VGG. We observe important differences across the networks and CNN layers for different scene factors and object categories. Finally, we demonstrate that our analysis based on computer-generated imagery translates to the network representation of natural images.