ICCV 2021 论文列表

General face recognition has seen remarkable progress in recent years. However, large age gap still remains a big challenge due to significant alterations in facial appearance and bone structure. Disentanglement plays a key role in partitioning face representations into identity-dependent and age-dependent components for age-invariant face recognition (AIFR). In this paper we propose a multi-task learning framework based on mutual information minimization (MT-MIM), which casts the disentangled representation learning as an objective of information constraints. The method trains a disentanglement network to minimize mutual information between the identity component and age component of the face image from the same person, and reduce the effect of age variations during the identification process. For quantitative measure of the degree of disentanglement, we verify that mutual information can represent as metric. The resulting identity-dependent representations are used for age-invariant face recognition. We evaluate MT-MIM on popular public-domain face aging datasets (FG-NET, MORPH Album 2, CACD and AgeDB) and obtained significant improvements over previous state-of-the-art methods. Specifically, our method exceeds the baseline models by over 0.4% on MORPH Album 2, and over 0.7% on CACD subsets, which are impressive improvements at the high accuracy levels of above 99% and an average of 94%.

A lifespan face synthesis (LFS) model aims to generate a set of photo-realistic face images of a person's whole life, given only one snapshot as reference. The generated face image given a target age code is expected to be age-sensitive reflected by bio-plausible transformations of shape and texture, while being identity preserving. This is extremely challenging because the shape and texture characteristics of a face undergo separate and highly nonlinear transformations w.r.t. age. Most recent LFS models are based on generative adversarial networks (GANs) whereby age code conditional transformations are applied to a latent face representation. They benefit greatly from the recent advancements of GANs. However, without explicitly disentangling their latent representations into the texture, shape and identity factors, they are fundamentally limited in modeling the nonlinear age-related transformation on texture and shape whilst preserving identity. In this work, a novel LFS model is proposed to disentangle the key face characteristics including shape, texture and identity so that the unique shape and texture age transformations can be modeled effectively. This is achieved by extracting shape, texture and identity features separately from an encoder. Critically, two transformation modules, one conditional convolution based and the other channel attention based, are designed for modeling the nonlinear shape and texture feature transformations respectively. This is to accommodate their rather distinct aging processes and ensure that our synthesized images are both age-sensitive and identity preserving. Extensive experiments show that our LFS model is clearly superior to the state-of-the-art alternatives.

Aiming at discovering and locating most distinctive objects from visual scenes, salient object detection (SOD) plays an essential role in various computer vision systems. Coming to the era of high resolution, SOD methods are facing new challenges. The major limitation of previous methods is that they try to identify the salient regions and estimate the accurate objects boundaries simultaneously with a single regression task at low-resolution. This practice ignores the inherent difference between the two difficult problems, resulting in poor detection quality. In this paper, we propose a novel deep learning framework for high-resolution SOD task, which disentangles the task into a low-resolution saliency classification network (LRSCN) and a high-resolution refinement network (HRRN). As a pixel-wise classification task, LRSCN is designed to capture sufficient semantics at low-resolution to identify the definite salient, background and uncertain image regions. HRRN is a regression task, which aims at accurately refining the saliency value of pixels in the uncertain region to preserve a clear object boundary at high-resolution with limited GPU memory. It is worth noting that by introducing uncertainty into the training process, our HRRN can well address the high-resolution refinement task without using any high-resolution training data. Extensive experiments on high-resolution saliency datasets as well as some widely used saliency benchmarks show that the proposed method achieves superior performance compared to the state-of-the-art methods.

Multi-label zero-shot learning (ZSL) is a more realistic counter-part of standard single-label ZSL since several objects can co-exist in a natural image. However, the occurrence of multiple objects complicates the reasoning and requires region-specific processing of visual features to preserve their contextual cues. We note that the best existing multi-label ZSL method takes a shared approach towards attending to region features with a common set of attention maps for all the classes. Such shared maps lead to diffused attention, which does not discriminatively focus on relevant locations when the number of classes are large. Moreover, mapping spatially-pooled visual features to the class semantics leads to inter-class feature entanglement, thus hampering the classification. Here, we propose an alternate approach towards region-based discriminability-preserving multi-label zero-shot classification. Our approach maintains the spatial resolution to preserve region-level characteristics and utilizes a bi-level attention module (BiAM) to enrich the features by incorporating both region and scene context information. The enriched region-level features are then mapped to the class semantics and only their class predictions are spatially pooled to obtain image-level predictions, thereby keeping the multi-class features disentangled. Our approach sets a new state of the art on two large-scale multi-label zero-shot benchmarks: NUS-WIDE and Open Images. On NUS-WIDE, our approach achieves an absolute gain of 6.9% mAP for ZSL, compared to the best published results.

In this work, we study the problem of human-object interaction (HOI) detection with large vocabulary object categories. Previous HOI studies are mainly conducted in the regime of limit object categories (e.g., 80 categories). Their solutions may face new difficulties in both object detection and interaction classification due to the increasing diversity of objects (e.g., 1000 categories). Different from previous methods, we formulate the HOI detection as a query problem. We propose a unified model to jointly discover the target objects and predict the corresponding interactions based on the human queries, thereby eliminating the need of using generic object detectors, extra steps to associate human-object instances, and multi-stream interaction recognition. This is achieved by a repurposed Transformer unit and a novel cascade detection over multi-scale feature maps. We observe that such a highly-coupled solution brings benefits for both object detection and interaction classification in a large vocabulary setting. To study the new challenges of the large vocabulary HOI detection, we assemble two datasets from the publicly available SWiG and 100 Days of Hands datasets. Experiments on these datasets validate that our proposed method can achieve a notable mAP improvement on HOI detection with a faster inference speed than existing one-stage HOI detectors.

Reasoning 3D shapes from 2D images is an essential yet challenging task, especially when only single-view images are at our disposal. While an object can have a complicated shape, individual parts are usually close to geometric primitives and thus are easier to model. Furthermore, parts provide a mid-level representation that is robust to appearance variations across objects in a particular category. In this work, we tackle the problem of 3D part discovery from only 2D image collections. Instead of relying on manually annotated parts for supervision, we propose a self-supervised approach, latent part discovery (LPD). Our key insight is to learn a novel part shape prior that allows each part to fit an object shape faithfully while constrained to have simple geometry. Extensive experiments on the synthetic ShapeNet, PartNet, and real-world Pascal 3D+ datasets show that our method discovers consistent object parts and achieves favorable reconstruction accuracy compared to the existing methods with the same level of supervision. Our project page with code is at https://chhankyao.github.io/lpd/.

Recent works find that AI algorithms learn biases from data. Therefore, it is urgent and vital to identify biases in AI algorithms. However, the previous bias identification pipeline overly relies on human experts to conjecture potential biases (e.g., gender), which may neglect other underlying biases not realized by humans. To help human experts better find the AI algorithms' biases, we study a new problem in this work -- for a classifier that predicts a target attribute of the input image, discover its unknown biased attribute. To solve this challenging problem, we use a hyperplane in the generative model's latent space to represent an image attribute; thus, the original problem is transformed to optimizing the hyperplane's normal vector and offset. We propose a novel total-variation loss within this framework as the objective function and a new orthogonalization penalty as a constraint. The latter prevents trivial solutions in which the discovered biased attribute is identical with the target or one of the known-biased attributes. Extensive experiments on both disentanglement datasets and real-world datasets show that our method can discover biased attributes and achieve better disentanglement w.r.t. target attributes. Furthermore, the qualitative results show that our method can discover unnoticeable biased attributes for various object and scene classifiers, proving our method's generalizability for detecting biased attributes in diverse domains of images.

We introduce DiscoBox, a novel framework that jointly learns instance segmentation and semantic correspondence using bounding box supervision. Specifically, we propose a self-ensembling framework where instance segmentation and semantic correspondence are jointly guided by a structured teacher in addition to the bounding box supervision. The teacher is a structured energy model incorporating a pairwise potential and a cross-image potential to model the pairwise pixel relationships both within and across the boxes. Minimizing the teacher energy simultaneously yields refined object masks and dense correspondences between intra-class objects, which are taken as pseudo-labels to supervise the task network and provide positive/negative correspondence pairs for dense contrastive learning. We show a symbiotic relationship where the two tasks mutually benefit from each other. Our best model achieves 37.9% AP on COCO instance segmentation, surpassing prior weakly supervised methods and is competitive to supervised methods. We also obtain state of the art weakly supervised results on PASCAL VOC12 and PF-PASCAL with real-time inference.

Disentangling data into interpretable and independent factors is critical for controllable generation tasks. With the availability of labeled data, supervision can help enforce the separation of specific factors as expected. However, it is often expensive or even impossible to label every single factor to achieve fully-supervised disentanglement. In this paper, we adopt a general setting where all factors that are hard to label or identify are encapsulated as a single unknown factor. Under this setting, we propose a flexible weakly-supervised multi-factor disentanglement framework DisUnknown, which Distills Unknown factors for enabling multi-conditional generation regarding both labeled and unknown factors. Specifically, a two-stage training approach is adopted to first disentangle the unknown factor with an effective and robust training method, and then train the final generator with the proper disentanglement of all labeled factors utilizing the unknown distillation. To demonstrate the generalization capacity and scalability of our method, we evaluate it on multiple benchmark datasets qualitatively and quantitatively and further apply it to various real-world applications on complicated datasets.

Data augmentation has been an indispensable tool to improve the performance of deep neural networks, however the augmentation can hardly transfer among different tasks and datasets. Consequently, a recent trend is to adopt AutoML technique to learn proper augmentation policy without extensive hand-crafted tuning. In this paper, we propose an efficient differentiable search algorithm called Direct Differentiable Augmentation Search (DDAS). It utilizes meta-learning with one-step gradient update and continuous relaxation to the expected training loss for efficient search. Our DDAS could achieve efficient augmentation search without approximations such as Gumbel-Softmax or second order gradient approximation. To further reduce the adverse effect of improper augmentations, we organize the search space into a two level hierarchy, in which we first decide whether to apply augmentation, and then determine the specific augmentation policy. On standard image classification benchmarks, our DDAS achieves state-of-the-art performance and efficiency tradeoff while reducing the search cost dramatically, e.g. 0.15 GPU hours for CIFAR-10. In addition, we also use DDAS to search augmentation for object detection task and achieve comparable performance with AutoAugment, while being 1000x faster. Code will be released in https://github.com/zxcvfd13502/DDAS_code.

Self-supervised Multi-view stereo (MVS) with a pretext task of image reconstruction has achieved significant progress recently. However, previous methods are built upon intuitions, lacking comprehensive explanations about the effectiveness of the pretext task in self-supervised MVS. To this end, we propose to estimate epistemic uncertainty in self-supervised MVS, accounting for what the model ignores. Specially, the limitations can be resorted into two folds: ambiguious supervision in foreground and noisy disturbance in background. To address these issues, we propose a novel Uncertainty reduction Multi-view Stereo (U-MVS) framework for self-supervised learning. To alleviate ambiguous supervision in foreground, we involve extra correspondence prior with a flow-depth consistency loss. The dense 2D correspondence of optical flows is used to regularize the 3D stereo correspondence in MVS. To handle the noisy disturbance in background, we use Monte-Carlo Dropout to acquire the uncertainty map and further filter the unreliable supervision signals on invalid regions. Extensive experiments on DTU and Tank&Temples benchmark show that our U-MVS framework achieves the best performance among unsupervised MVS methods, with competitive performance with its supervised opponents.

We present a method for differentiable rendering of 3D surfaces that supports both explicit and implicit representations, provides derivatives at occlusion boundaries, and is fast and simple to implement. The method first samples the surface using non-differentiable rasterization, then applies differentiable, depth-aware point splatting to produce the final image. Our approach requires no differentiable meshing or rasterization steps, making it efficient for large 3D models and applicable to isosurfaces extracted from implicit surface definitions. We demonstrate the effectiveness of our method for implicit-, mesh-, and parametric-surface-based inverse rendering and neural-network training applications. In particular, we show for the first time efficient, differentiable rendering of an isosurface extracted from a neural radiance field (NeRF), and demonstrate surface-based, rather than volume-based, rendering of a NeRF.

A standard practice of deploying deep neural networks is to apply the same architecture to all the input instances. However, a fixed architecture may not be suitable for different data with high diversity. To boost the model capacity, existing methods usually employ larger convolutional kernels or deeper network layers, which incurs prohibitive computational costs. In this paper, we address this issue by proposing Differentiable Dynamic Wirings (DDW), which learns the instance-aware connectivity that creates different wiring patterns for different instances. 1) Specifically, the network is initialized as a complete directed acyclic graph, where the nodes represent convolutional blocks and the edges represent the connection paths. 2) We generate edge weights by a learnable module, Router, and select the edges whose weights are larger than a threshold, to adjust the connectivity of the neural network structure. 3) Instead of using the same path of the network, DDW aggregates features dynamically in each node, which allows the network to have more representation power. To facilitate effective training, we further represent the network connectivity of each sample as an adjacency matrix. The matrix is updated to aggregate features in the forward pass, cached in the memory, and used for gradient computing in the backward pass. We validate the effectiveness of our approach with several mainstream architectures, including MobileNetV2, ResNet, ResNeXt, and RegNet. Extensive experiments are performed on ImageNet classification and COCO object detection, which demonstrates the effectiveness and generalization ability of our approach.

Exploiting convolutional neural networks for point cloud processing is quite challenging, due to the inherent irregular distribution and discrete shape representation of point clouds. To address these problems, many handcrafted convolution variants have sprung up in recent years. Though with elaborate design, these variants could be far from optimal in sufficiently capturing diverse shapes formed by discrete points. In this paper, we propose PointSeaConv, i.e., a novel differential convolution search paradigm on point clouds. It can work in a purely data-driven manner and thus is capable of auto-creating a group of suitable convolutions for geometric shape modeling. We also propose a joint optimization framework for simultaneous search of internal convolution and external architecture, and introduce epsilon-greedy algorithm to alleviate the effect of discretization error. As a result, PointSeaNet, a deep network that is sufficient to capture geometric shapes at both convolution level and architecture level, can be searched out for point cloud processing. Extensive experiments strongly evidence that our proposed PointSeaNet surpasses current handcrafted deep models on challenging benchmarks across multiple tasks with remarkable margins.

One of the important research topics in image generative models is to disentangle the spatial contents and styles for their separate control. Although StyleGAN can generate content feature vectors from random noises, the resulting spatial content control is primarily intended for minor spatial variations, and the disentanglement of global content and styles is by no means complete. Inspired by a mathematical understanding of normalization and attention, here we present a novel hierarchical adaptive Diagonal spatial ATtention (DAT) layers to separately manipulate the spatial contents from styles in a hierarchical manner. Using DAT and AdaIN, our method enables coarse-to-fine level disentanglement of spatial contents and styles. In addition, our generator can be easily integrated into the GAN inversion framework so that the content and style of translated images from multi-domain image translation tasks can be flexibly controlled. By using various datasets, we confirm that the proposed method not only outperforms the existing models in disentanglement scores, but also provides more flexible control over spatial features in the generated images.

Common deep neural networks (DNNs) for image classification have been shown to rely on shortcut opportunities (SO) in the form of predictive and easy-to-represent visual factors. This is known as shortcut learning and leads to impaired generalization. In this work, we show that common DNNs also suffer from shortcut learning when predicting only basic visual object factors of variation (FoV) such as shape, color, or texture. We argue that besides shortcut opportunities, generalization opportunities (GO) are also an inherent part of real-world vision data and arise from partial independence between predicted classes and FoVs. We also argue that it is necessary for DNNs to exploit GO to overcome shortcut learning. Our core contribution is to introduce the Diagnostic Vision Benchmark suite DiagViB-6, which includes datasets and metrics to study a network's shortcut vulnerability and generalization capability for six independent FoV. In particular, DiagViB-6 allows controlling the type and degree of SO and GO in a dataset. We benchmark a wide range of popular vision architectures and show that they can exploit GO only to a limited extent.

Nowadays, there is an abundance of data involving images and surrounding free-form text weakly corresponding to those images. Weakly Supervised phrase-Grounding (WSG) deals with the task of using this data to learn to localize (or to ground) arbitrary text phrases in images without any additional annotations. However, most recent SotA methods for WSG assume an existence of a pre-trained object detector, relying on it to produce the ROIs for localization. In this work, we focus on the task of Detector-Free WSG (DF-WSG) to solve WSG without relying on a pre-trained detector. We directly learn everything from the images and associated free-form text pairs, thus potentially gaining advantage on the categories unsupported by the detector. The key idea behind our proposed Grounding by Separation (GbS) method is synthesizing `text to image-regions' associations by random alpha-blending of arbitrary image pairs and using the corresponding texts of the pair as conditions to recover the alpha map from the blended image via a segmentation network. At test time, this allows using the query phrase as a condition for a non-blended query image, thus interpreting the test image as a composition of a region corresponding to the phrase and the complement region. Using this approach we demonstrate a significant accuracy improvement, up to 8.5% over previous DF-WSG SotA, for a range of benchmarks including Flickr30K, Visual Genome, and ReferIt, as well as a significant complementary improvement (above 7%) over the detector-based approaches for WSG.

Advances in deep learning, combined with availability of large datasets, have led to impressive improvements in face presentation attack detection research. However, state of the art face antispoofing systems are still vulnerable to novel types of attacks that are never seen during training. Moreover, even if such attacks are correctly detected, these systems lack the ability to adapt to newly encountered attacks. The post-training ability of continually detecting new types of attacks and self-adaptation to identify these attack types, after the initial detection phase, is highly appealing. In this paper, we enable a deep neural network to detect anomalies in the observed input data points as potential new types of attacks by suppressing the confidence-level of the network outside the training samples' distribution. We then use experience replay to update the model to incorporate knowledge about new types of attacks without forgetting the past learned attack types. Experimental results are provided to demonstrate the effectiveness of the proposed method on the OULU and Idiap datasets as well as a newly introduced dataset, all of which exhibit a variety of attack types.

We propose a new approach to detect atypicality in persuasive imagery. Unlike atypicality which has been studied in prior work, persuasive atypicality has a particular purpose to convey meaning, and relies on understanding the common-sense spatial relations of objects. We propose a self-supervised attention-based technique which captures contextual compatibility, and models spatial relations in a precise manner. We further experiment with capturing common sense through the semantics of co-occurring object classes. We verify our approach on a dataset of atypicality in visual advertisements, as well as a second dataset capturing atypicality that has no persuasive intent.