Total: 1660
The canonical approach to video-and-language learning (e.g., video question answering) dictates a neural model to learn from offline-extracted dense video features from vision models and text features from language models. These feature extractors are trained independently and usually on tasks different from the target domains, rendering these fixed features sub-optimal for downstream tasks. Moreover, due to the high computational overload of dense video features, it is often difficult (or infeasible) to plug feature extractors directly into existing approaches for easy finetuning. To provide a remedy to this dilemma, we propose a generic framework CLIPBERT that enables affordable end-to-end learning for video-and-language tasks, by employing sparse sampling, where only a single or a few sparsely sampled short clips from a video are used at each training step. Experiments on text-to-video retrieval and video question answering on six datasets demonstrate that CLIPBERT outperforms (or is on par with) existing methods that exploit full-length videos, suggesting that end-to-end learning with just a few sparsely sampled clips is often more accurate than using densely extracted offline features from full-length videos, proving the proverbial less-is-more principle. Videos in the datasets are from considerably different domains and lengths, ranging from 3-second generic-domain GIF videos to 180-second YouTube human activity videos, showing the generalization ability of our approach. Comprehensive ablation studies and thorough analyses are provided to dissect what factors lead to this success.
We introduce a real-time, high-resolution background replacement technique which operates at 30fps in 4K resolution, and 60fps for HD on a modern GPU. Our technique is based on background matting, where an additional frame of the background is captured and used to inform the alpha matte and the foreground layer. The main challenge is to compute a high-quality alpha matte, preserving strand-level hair details, while processing high-resolution images in real-time. To achieve this goal, we employ two neural networks; the base network computes a low-resolution result which is refined by a second network operating at high-resolution on selective patches. We introduce two large-scale video and image matting datasets: VideoMatte240K and PhotoMatte13K/85. Our approach yields higher quality results compared to the previous state-of-the-art in background matting, while simultaneously yielding a dramatic boost in both speed and resolution.
Specialized domain knowledge is often necessary to accurately annotate training sets for in-depth analysis, but can be burdensome and time-consuming to acquire from domain experts. This issue arises prominently in automated behavior analysis, in which agent movements or actions of interest are detected from video tracking data. To reduce annotation effort, we present TREBA: a method to learn annotation-sample efficient trajectory embedding for behavior analysis, based on multi-task self-supervised learning. The tasks in our method can be efficiently engineered by domain experts through a process we call "task programming", which uses programs to explicitly encode structured knowledge from domain experts. Total domain expert effort can be reduced by exchanging data annotation time for the construction of a small number of programmed tasks. We evaluate this trade-off using data from behavioral neuroscience, in which specialized domain knowledge is used to identify behaviors. We present experimental results in three datasets across two domains: mice and fruit flies. Using embeddings from TREBA, we reduce annotation burden by up to a factor of 10 without compromising accuracy compared to state-of-the-art features. Our results thus suggest that task programming and self-supervision can be an effective way to reduce annotation effort for domain experts.
Siamese networks have become a common structure in various recent models for unsupervised visual representation learning. These models maximize the similarity between two augmentations of one image, subject to certain conditions for avoiding collapsing solutions. In this paper, we report surprising empirical results that simple Siamese networks can learn meaningful representations even using none of the following: (i) negative sample pairs, (ii) large batches, (iii) momentum encoders. Our experiments show that collapsing solutions do exist for the loss and structure, but a stop-gradient operation plays an essential role in preventing collapsing. We provide a hypothesis on the implication of stop-gradient, and further show proof-of-concept experiments verifying it. Our "SimSiam" method achieves competitive results on ImageNet and downstream tasks. We hope this simple baseline will motivate people to rethink the roles of Siamese architectures for unsupervised representation learning. Code is made available. (https://github.com/facebookresearch/simsiam)
A key challenge of learning the geometry of dressed humans lies in the limited availability of the ground truth data (e.g., 3D scanned models), which results in the performance degradation of 3D human reconstruction when applying to real world imagery. We address this challenge by leveraging a new data resource: a number of social media dance videos that span diverse appearance, clothing styles, performances, and identities. Each video depicts dynamic movements of the body and clothes of a single person while lacking the 3D ground truth geometry. To utilize these videos, we present a new method to use the local transformation that warps the predicted local geometry of the person from an image to that of the other image at a different time instant. With the transformation, the predicted geometry can be self-supervised by the warped geometry from the other image. In addition, we jointly learn the depth along with the surface normals, which are highly responsive to local texture, wrinkle, and shade by maximizing their geometric consistency. Our method is end-to-end trainable, resulting in high fidelity depth estimation that predicts fine geometry faithful to the input real image. We demonstrate that our method outperforms the state-of-the-art human depth estimation and human shape recovery approaches on both real and rendered images.
Deep generative models allow for photorealistic image synthesis at high resolutions. But for many applications, this is not enough: content creation also needs to be controllable. While several recent works investigate how to disentangle underlying factors of variation in the data, most of them operate in 2D and hence ignore that our world is three-dimensional. Further, only few works consider the compositional nature of scenes. Our key hypothesis is that incorporating a compositional 3D scene representation into the generative model leads to more controllable image synthesis. Representing scenes as compositional generative neural feature fields allows us to disentangle one or multiple objects from the background as well as individual objects' shapes and appearances while learning from unstructured and unposed image collections without any additional supervision. Combining this scene representation with a neural rendering pipeline yields a fast and realistic image synthesis model. As evidenced by our experiments, our model is able to disentangle individual objects and allows for translating and rotating them in the scene as well as changing the camera pose.
Invertible networks have various benefits for image denoising since they are lightweight, information-lossless, and memory-saving during back-propagation. However, applying invertible models to remove noise is challenging because the input is noisy, and the reversed output is clean, following two different distributions. We propose an invertible denoising network, InvDN, to address this challenge. InvDN transforms the noisy input into a low-resolution clean image and a latent representation containing noise. To discard noise and restore the clean image, InvDN replaces the noisy latent representation with another one sampled from a prior distribution during reversion. The denoising performance of InvDN is better than all the existing competitive models, achieving a new state-of-the-art result for the SIDD dataset while enjoying less run time. Moreover, the size of InvDN is far smaller, only having 4.2% of the number of parameters compared to the most recently proposed DANet. Further, via manipulating the noisy latent representation, InvDN is also able to generate noise more similar to the original one. Our code is available at: https://github.com/Yang-Liu1082/InvDN.git.
A video prediction model that generalizes to diverse scenes would enable intelligent agents such as robots to perform a variety of tasks via planning with the model. However, while existing video prediction models have produced promising results on small datasets, they suffer from severe underfitting when trained on large and diverse datasets. To address this underfitting challenge, we first observe that the ability to train larger video prediction models is often bottlenecked by the memory constraints of GPUs or TPUs. In parallel, deep hierarchical latent variable models can produce higher quality predictions by capturing the multi-level stochasticity of future observations, but end-to-end optimization of such models is notably difficult. Our key insight is that greedy and modular optimization of hierarchical autoencoders can simultaneously address both the memory constraints and the optimization challenges of large-scale video prediction. We introduce Greedy Hierarchical Variational Autoencoders (GHVAEs), a method that learns high-fidelity video predictions by greedily training each level of a hierarchical autoencoder. In comparison to state-of-the-art models, GHVAEs provide 17-55% gains in prediction performance on four video datasets, a 35-40% higher success rate on real robot tasks, and can improve performance monotonically by simply adding more modules.
Deep neural networks for video classification, just like image classification networks, may be subjected to adversarial manipulation. The main difference between image classifiers and video classifiers is that the latter usually use temporal information contained within the video. In this work we present a manipulation scheme for fooling video classifiers by introducing a flickering temporal perturbation that in some cases may be unnoticeable by human observers and is implementable in the real world. After demonstrating the manipulation of action classification of single videos, we generalize the procedure to make universal adversarial perturbation, achieving high fooling ratio. In addition, we generalize the universal perturbation and produce a temporal-invariant perturbation, which can be applied to the video without synchronizing the perturbation to the input. The attack was implemented on several target models and the transferability of the attack was demonstrated. These properties allow us to bridge the gap between simulated environment and real-world application, as will be demonstrated in this paper for the first time for an over-the-air flickering attack.
Recently, referring image segmentation has aroused widespread interest. Previous methods perform the multi-modal fusion between language and vision at the decoding side of the network. And, linguistic feature interacts with visual feature of each scale separately, which ignores the continuous guidance of language to multi-scale visual features. In this work, we propose an encoder fusion network (EFN), which transforms the visual encoder into a multi-modal feature learning network, and uses language to refine the multi-modal features progressively. Moreover, a co-attention mechanism is embedded in the EFN to realize the parallel update of multi-modal features, which can promote the consistent of the cross-modal information representation in the semantic space. Finally, we propose a boundary enhancement module (BEM) to make the network pay more attention to the fine structure. The experiment results on four benchmark datasets demonstrate that the proposed approach achieves the state-of-the-art performance under different evaluation metrics without any post-processing.
Active stereo cameras that recover depth from structured light captures have become a cornerstone sensor modality for 3D scene reconstruction and understanding tasks across application domains. Active stereo cameras project a pseudo-random dot pattern on object surfaces to extract disparity independently of object texture. Such hand-crafted patterns are designed in isolation from the scene statistics, ambient illumination conditions, and the reconstruction method. In this work, we propose a method to jointly learn structured illumination and reconstruction, parameterized by a diffractive optical element and a neural network, in an end-to-end fashion. To this end, we introduce a differentiable image formation model for active stereo, relying on both wave and geometric optics, and a trinocular reconstruction network. The jointly optimized pattern, which we dub "Polka Lines," together with the reconstruction network, makes accurate active-stereo depth estimates across imaging conditions. We validate the proposed method in simulation and using with an experimental prototype, and we demonstrate several variants of the Polka Lines patterns specialized to the illumination conditions.
Although recent inpainting approaches have demonstrated significant improvement with deep neural networks, they still suffer from artifacts such as blunt structures and abrupt colors when filling in the missing regions. To address these issues, we propose an external-internal inpainting scheme with a monochromic bottleneck that helps image inpainting models remove these artifacts. In the external learning stage, we reconstruct missing structures and details in the monochromic space to reduce the learning dimension. In the internal learning stage, we propose a novel internal color propagation method with progressive learning strategies for consistent color restoration. Extensive experiments demonstrate that our proposed scheme helps image inpainting models produce more structure-preserved and visually compelling results.
The classical matching pipeline used for visual localization typically involves three steps: (i) local feature detection and description, (ii) feature matching, and (iii) outlier rejection. Recently emerged correspondence networks propose to perform those steps inside a single network but suffer from low matching resolution due to the memory bottleneck. In this work, we propose a new perspective to estimate correspondences in a detect-to-refine manner, where we first predict patch-level match proposals and then refine them. We present Patch2Pix, a novel refinement network that refines match proposals by regressing pixel-level matches from the local regions defined by those proposals and jointly rejecting outlier matches with confidence scores. Patch2Pix is weakly supervised to learn correspondences that are consistent with the epipolar geometry of an input image pair. We show that our refinement network significantly improves the performance of correspondence networks on image matching, homography estimation, and localization tasks. In addition, we show that our learned refinement generalizes to fully-supervised methods without re-training, which leads us to state-of-the-art localization performance. The code is available at https://github.com/GrumpyZhou/patch2pix.
Occluded person re-identification (Re-ID) is a challenging task as persons are frequently occluded by various obstacles or other persons, especially in the crowd scenario. To address these issues, we propose a novel end-to-end Part-Aware Transformer (PAT) for occluded person Re-ID through diverse part discovery via a transformer encoder-decoder architecture, including a pixel context based transformer encoder and a part prototype based transformer decoder. The proposed PAT model enjoys several merits. First, to the best of our knowledge, this is the first work to exploit the transformer encoder-decoder architecture for occluded person Re-ID in a unified deep model. Second, to learn part prototypes well with only identity labels, we design two effective mechanisms including part diversity and part discriminability. Consequently, we can achieve diverse part discovery for occluded person Re-ID in a weakly supervised manner. Extensive experimental results on six challenging benchmarks for three tasks (occluded, partial and holistic Re-ID) demonstrate that our proposed PAT performs favorably against stat-of-the-art methods.
We present a novel counterfactual framework for both Zero-Shot Learning (ZSL) and Open-Set Recognition (OSR), whose common challenge is generalizing to the unseen-classes by only training on the seen-classes. Our idea stems from the observation that the generated samples for unseen-classes are often out of the true distribution, which causes severe recognition rate imbalance between the seen-class (high) and unseen-class (low). We show that the key reason is that the generation is not Counterfactual Faithful, and thus we propose a faithful one, whose generation is from the sample-specific counterfactual question: What would the sample look like, if we set its class attribute to a certain class, while keeping its sample attribute unchanged? Thanks to the faithfulness, we can apply the Consistency Rule to perform unseen/seen binary classification, by asking: Would its counterfactual still look like itself? If "yes", the sample is from a certain class, and "no" otherwise. Through extensive experiments on ZSL and OSR, we demonstrate that our framework effectively mitigates the seen/unseen imbalance and hence significantly improves the overall performance. Note that this framework is orthogonal to existing methods, thus, it can serve as a new baseline to evaluate how ZSL/OSR models generalize. Codes are available at https://github.com/yue-zhongqi/gcm-cf.
Recently, person re-identification (ReID) has vastly benefited from the surging waves of data-driven methods. However, these methods are still not reliable enough for real-world deployments, due to the insufficient generalization capability of the models learned on existing benchmarks that have limitations in multiple aspects, including limited data scale, capture condition variations, and appearance diversities. To this end, we collect a new dataset named Person30K with the following distinct features: 1) a very large scale containing 1.38 million images of 30K identities, 2) a large capture system containing 6,497 cameras deployed at 89 different sites, 3) abundant sample diversities including varied backgrounds and diverse person poses. Furthermore, we propose a domain generalization ReID method, dual-meta generalization network (DMG-Net), to exploit the merits of meta-learning in both the training procedure and the metric space learning. Concretely, we design a "learning then generalization evaluation" meta-training procedure and a meta-discrimination loss to enhance model generalization and discrimination capabilities. Comprehensive experiments validate the effectiveness of our DMG-Net. (Dataset and code will be released.)
Visual Place Recognition is a challenging task for robotics and autonomous systems, which must deal with the twin problems of appearance and viewpoint change in an always changing world. This paper introduces Patch-NetVLAD, which provides a novel formulation for combining the advantages of both local and global descriptor methods by deriving patch-level features from NetVLAD residuals. Unlike the fixed spatial neighborhood regime of existing local keypoint features, our method enables aggregation and matching of deep-learned local features defined over the feature-space grid. We further introduce a multi-scale fusion of patch features that have complementary scales (i.e. patch sizes) via an integral feature space and show that the fused features are highly invariant to both condition (season, structure, and illumination) and viewpoint (translation and rotation) changes. Patch-NetVLAD achieves state-of-the-art visual place recognition results in computationally limited scenarios, validated on a range of challenging real-world datasets, including winning the Facebook Mapillary Visual Place Recognition Challenge at ECCV2020. It is also adaptable to user requirements, with a speed-optimised version operating over an order of magnitude faster than the state-of-the-art. By combining superior performance with improved computational efficiency in a configurable framework, Patch-NetVLAD is well suited to enhance both stand-alone place recognition capabilities and the overall performance of SLAM systems.
Stereophonic audio, especially binaural audio, plays an essential role in immersive viewing environments. Recent research has explored generating stereophonic audios guided by visual cues and multi-channel audio collections in a fully-supervised manner. However, due to the requirement of professional recording devices, existing datasets are limited in scale and variety, which impedes the generalization of supervised methods to real-world scenarios. In this work, we propose PseudoBinaural, an effective pipeline that is free of binaural recordings. The key insight is to carefully build pseudo visual-stereo pairs with mono data for training. Specifically, we leverage spherical harmonic decomposition and head-related impulse response (HRIR) to identify the relationship between the location of a sound source and the received binaural audio. Then in the visual modality, corresponding visual cues of the mono data are manually placed at sound source positions to form the pairs. Compared to fully-supervised paradigms, our binaural-recording-free pipeline shows great stability in the cross-dataset evaluation and comparable performance under subjective preference. Moreover, combined with binaural recorded data, our method is able to further boost the performance of binaural audio generation under supervised settings.
Gaze target detection aims to infer where each person in a scene is looking. Existing works focus on 2D gaze and 2D saliency, but fail to exploit 3D contexts. In this work, we propose a three-stage method to simulate the human gaze inference behavior in 3D space. In the first stage, we introduce a coarse-to-fine strategy to robustly estimate a 3D gaze orientation from the head. The predicted gaze is decomposed into a planar gaze on the image plane and a depth-channel gaze. In the second stage, we develop a Dual Attention Module (DAM), which takes the planar gaze to produce the filed of view and masks interfering objects regulated by depth information according to the depth-channel gaze. In the third stage, we use the generated dual attention as guidance to perform two sub-tasks: (1) identifying whether the gaze target is inside or out of the image; (2) locating the target if inside. Extensive experiments demonstrate that our approach performs favorably against state-of-the-art methods on GazeFollow and VideoAttentionTarget datasets.
Recent works on localization and mapping from privacy preserving line features have made significant progress towards addressing the privacy concerns arising from cloud-based solutions in mixed reality and robotics. The requirement for calibrated cameras is a fundamental limitation for these approaches, which prevents their application in many crowd-sourced mapping scenarios. In this paper, we propose a solution to the uncalibrated privacy preserving localization and mapping problem. Our approach simultaneously recovers the intrinsic and extrinsic calibration of a camera from line-features only. This enables uncalibrated devices to both localize themselves within an existing map as well as contribute to the map, while preserving the privacy of the image contents. Furthermore, we also derive a solution to bootstrapping maps from scratch using only uncalibrated devices. Our approach provides comparable performance to the calibrated scenario and the privacy compromising alternatives based on traditional point features.
In medical image analysis, it is typical to collect multiple annotations, each from a different clinical expert or rater, in the expectation that possible diagnostic errors could be mitigated. Meanwhile, from the computer vision practitioner viewpoint, it has been a common practice to adopt the ground-truth obtained via either the majority-vote or simply one annotation from a preferred rater. This process, however, tends to overlook the rich information of agreement or disagreement ingrained in the raw multi-rater annotations. To address this issue, we propose to explicitly model the multi-rater (dis-)agreement, dubbed MRNet, which has two main contributions. First, an expertise-aware inferring module or EIM is devised to embed the expertise level of individual raters as prior knowledge, to form high-level semantic features. Second, our approach is capable of reconstructing multi-rater gradings from coarse predictions, with the multi-rater (dis-)agreement cues being further exploited to improve the segmentation performance. To our knowledge, our work is the first in producing calibrated predictions under different expertise levels for medical image segmentation. Extensive empirical experiments are conducted across five medical segmentation tasks of diverse imaging modalities. In these experiments, superior performance of our MRNet is observed comparing to the state-of-the-arts, indicating the effectiveness and applicability of our MRNet toward a wide range of medical segmentation tasks.
We propose a novel point annotated setting for the weakly semi-supervised object detection task, in which the dataset comprises small fully annotated images and large weakly annotated images by points. It achieves a balance between tremendous annotation burden and detection performance. Based on this setting, we analyze existing detectors and find that these detectors have difficulty in fully exploiting the power of the annotated points. To solve this, we introduce a new detector, Point DETR, which extends DETR by adding a point encoder. Extensive experiments conducted on MS-COCO dataset in various data settings show the effectiveness of our method. In particular, when using 20% fully labeled data from COCO, our detector achieves a promising performance, 33.3 AP, which outperforms a strong baseline (FCOS) by 2.0 AP, and we demonstrate the point annotations bring over 10 points in various AR metrics.
Recent development of Under-Display Camera (UDC) systems provides a true bezel-less and notch-free viewing experience on smartphones (and TV, laptops, tablets), while allowing images to be captured from the selfie camera embedded underneath. In a typical UDC system, the microstructure of the semi-transparent organic light-emitting diode (OLED) pixel array attenuates and diffracts the incident light on the camera, resulting in significant image quality degradation. Oftentimes, noise, flare, haze, and blur can be observed in UDC images. In this work, we aim to analyze and tackle the aforementioned degradation problems. We define a physics-based image formation model to better understand the degradation. In addition, we utilize one of the world's first commodity UDC smartphone prototypes to measure the real-world Point Spread Function (PSF) of the UDC system, and provide a model-based data synthesis pipeline to generate realistically degraded images. We specially design a new domain knowledge-enabled Dynamic Skip Connection Network (DISCNet) to restore the UDC images. We demonstrate the effectiveness of our method through extensive experiments on both synthetic and real UDC data. Our physics-based image formation model and proposed DISCNet can provide foundations for further exploration in UDC image restoration, and even for general diffraction artifact removal in a broader sense.
It is nontrivial to store rapidly growing big data nowadays, which demands high-performance lossless compression techniques. Likelihood-based generative models have witnessed their success on lossless compression, where flow based models are desirable in allowing exact data likelihood optimisation with bijective mappings. However, common continuous flows are in contradiction with the discreteness of coding schemes, which requires either 1) imposing strict constraints on flow models that degrades the performance or 2) coding numerous bijective mapping errors which reduces the efficiency. In this paper, we investigate volume preserving flows for lossless compression and show that a bijective mapping without error is possible. We propose Numerical Invertible Volume Preserving Flow (iVPF) which is derived from the general volume preserving flows. By introducing novel computation algorithms on flow models, an exact bijective mapping is achieved without any numerical error. We also propose a lossless compression algorithm based on iVPF. Experiments on various datasets show that the algorithm based on iVPF achieves state-of-the-art compression ratio over lightweight compression algorithms.