2D convolution (Conv2d), which is responsible for extracting features from the input image, is one of the key modules of a convolutional neural network (CNN). However, Conv2d is vulnerable to image corruptions and adversarial samples. It is an important yet rarely investigated problem that whether we can design a more robust alternative of Conv2d for more reliable feature extraction. In this paper, inspired by the recently developed learnable sparse transform that learns to convert the CNN features into a compact and sparse latent space, we design a novel building block, denoted by RConv-MK, to strengthen the robustness of extracted convolutional features. Our method leverages a set of learnable kernels of different sizes to extract features at different frequencies and employs a normalized soft thresholding operator to adaptively remove noises and trivial features at different corruption levels. Extensive experiments on clean images, corrupted images as well as adversarial samples validate the effectiveness of the proposed robust module for reliable visual recognition. The source codes are enclosed in the submission.
Video frame interpolation (VFI) aims to improve the temporal resolution of a video sequence. Most of the existing deep learning based VFI methods adopt off-the-shelf optical flow algorithms to estimate the bidirectional flows and interpolate the missing frames accordingly. Though having achieved a great success, these methods require much human experience to tune the bidirectional flows and often generate unpleasant results when the estimated flows are not accurate. In this work, we rethink the VFI problem and formulate it as a continuous image transition (CIT) task, whose key issue is to transition an image from one space to another space continuously. More specifically, we learn to implicitly decouple the images into a translatable flow space and a non-translatable feature space. The former depicts the translatable states between the given images, while the later aims to reconstruct the intermediate features that cannot be directly translated. In this way, we can easily perform image interpolation in the flow space and intermediate image synthesis in the feature space, obtaining a CIT model. The proposed space decoupled learning (SDL) approach is simple to implement, while it provides an effective framework to a variety of CIT problems beyond VFI, such as style transfer and image morphing. Our extensive experiments on a variety of CIT tasks demonstrate the superiority of SDL to existing methods. The source code and models can be found at \url{https://github.com/yangxy/SDL}.
Domain adaptive semantic segmentation aims to learn a model with the supervision of source domain data, and produce satisfactory dense predictions on unlabeled target domain. One popular solution to this challenging task is self-training, which selects high-scoring predictions on target samples as pseudo labels for training. However, the produced pseudo labels often contain much noise because the model is biased to source domain as well as majority categories. To address the above issues, we propose to directly explore the intrinsic pixel distributions of target domain data, instead of heavily relying on the source domain. Specifically, we simultaneously cluster pixels and rectify pseudo labels with the obtained cluster assignments. This process is done in an online fashion so that pseudo labels could co-evolve with the segmentation model without extra training rounds. To overcome the class imbalance problem on long-tailed categories, we employ a distribution alignment technique to enforce the marginal class distribution of cluster assignments to be close to that of pseudo labels. The proposed method, namely Class-balanced Pixel-level Self-Labeling (CPSL), improves the segmentation performance on target domain over state-of-the-arts by a large margin, especially on long-tailed categories.
Label assignment (LA), which aims to assign each training sample a positive (pos) and a negative (neg) loss weight, plays an important role in object detection. Existing LA methods mostly focus on the design of pos weighting function, while the neg weight is directly derived from the pos weight. Such a mechanism limits the learning capacity of detectors. In this paper, we explore a new weighting paradigm, termed dual weighting (DW), to specify pos and neg weights separately. We first identify the key influential factors of pos/neg weights by analyzing the evaluation metrics in object detection, and then design the pos and neg weighting functions based on them. Specifically, the pos weight of a sample is determined by the consistency degree between its classification and localization scores, while the neg weight is decomposed into two terms: the probability that it is a neg sample and its importance conditioned on being a neg sample. Such a weighting strategy offers greater flexibility to distinguish between important and less important samples, resulting in a more effective object detector. Equipped with the proposed DW method, a single FCOS-ResNet-50 detector can reach 41.5% mAP on COCO under 1x schedule, outperforming other existing LA methods. It consistently improves the baselines on COCO by a large margin under various backbones without bells and whistles. Code is available at https://github.com/strongwolf/DW.
Scene text image super-resolution aims to increase the resolution and readability of the text in low-resolution images. Though significant improvement has been achieved by deep convolutional neural networks (CNNs), it remains difficult to reconstruct high-resolution images for spatially deformed texts, especially rotated and curve-shaped ones. This is because the current CNN-based methods adopt locality-based operations, which are not effective to deal with the variation caused by deformations. In this paper, we propose a CNN based Text ATTention network (TATT) to address this problem. The semantics of the text are firstly extracted by a text recognition module as text prior information. Then we design a novel transformer-based module, which leverages global attention mechanism, to exert the semantic guidance of text prior to the text reconstruction process. In addition, we propose a text structure consistency loss to refine the visual appearance by imposing structural consistency on the reconstructions of regular and deformed texts. Experiments on the benchmark TextZoom dataset show that the proposed TATT not only achieves state-of-the-art performance in terms of PSNR/SSIM metrics, but also significantly improves the recognition accuracy in the downstream text recognition task, particularly for text instances with multi-orientation and curved shapes. Code is available at https://github.com/mjq11302010044/TATT.
Denoising and demosaicking are two essential steps to reconstruct a clean full-color image from the raw data. Recently, joint denoising and demosaicking (JDD) for burst images, namely JDD-B, has attracted much attention by using multiple raw images captured in a short time to reconstruct a single high-quality image. One key challenge of JDD-B lies in the robust alignment of image frames. State-of-the-art alignment methods in feature domain cannot effectively utilize the temporal information of burst images, where large shifts commonly exist due to camera and object motion. In addition, the higher resolution (e.g., 4K) of modern imaging devices results in larger displacement between frames. To address these challenges, we design a differentiable two-stage alignment scheme sequentially in patch and pixel level for effective JDD-B. The input burst images are firstly aligned in the patch level by using a differentiable progressive block matching method, which can estimate the offset between distant frames with small computational cost. Then we perform implicit pixel-wise alignment in full-resolution feature domain to refine the alignment results. The two stages are jointly trained in an end-to-end manner. Extensive experiments demonstrate the significant improvement of our method over existing JDD-B methods. Codes are available at https://github.com/GuoShi28/2StageAlign.
Single image super-resolution (SISR) with generative adversarial networks (GAN) has recently attracted increasing attention due to its potentials to generate rich details. However, the training of GAN is unstable, and it often introduces many perceptually unpleasant artifacts along with the generated details. In this paper, we demonstrate that it is possible to train a GAN-based SISR model which can stably generate perceptually realistic details while inhibiting visual artifacts. Based on the observation that the local statistics (e.g., residual variance) of artifact areas are often different from the areas of perceptually friendly details, we develop a framework to discriminate between GAN-generated artifacts and realistic details, and consequently generate an artifact map to regularize and stabilize the model training process. Our proposed locally discriminative learning (LDL) method is simple yet effective, which can be easily plugged in off-the-shelf SISR methods and boost their performance. Experiments demonstrate that LDL outperforms the state-of-the-art GAN based SISR methods, achieving not only higher reconstruction accuracy but also superior perceptual quality on both synthetic and real-world datasets. Codes and models are available at https://github.com/csjliang/LDL.
Recently, transformer-based methods have demonstrated impressive results in various vision tasks, including image super-resolution (SR), by exploiting the self-attention (SA) for feature extraction. However, the computation of SA in most existing transformer based models is very expensive, while some employed operations may be redundant for the SR task. This limits the range of SA computation and consequently the SR performance. In this work, we propose an efficient long-range attention network (ELAN) for image SR. Specifically, we first employ shift convolution (shift-conv) to effectively extract the image local structural information while maintaining the same level of complexity as 1x1 convolution, then propose a group-wise multi-scale self-attention (GMSA) module, which calculates SA on non-overlapped groups of features using different window sizes to exploit the long-range image dependency. A highly efficient long-range attention block (ELAB) is then built by simply cascading two shift-conv with a GMSA module, which is further accelerated by using a shared attention mechanism. Without bells and whistles, our ELAN follows a fairly simple design by sequentially cascading the ELABs. Extensive experiments demonstrate that ELAN obtains even better results against the transformer-based SR models but with significantly less complexity. The source code can be found at https://github.com/xindongzhang/ELAN.
Many video instance segmentation (VIS) methods partition a video sequence into individual frames to detect and segment objects frame by frame. However, such a frame-in frame-out (FiFo) pipeline is ineffective to exploit the temporal information. Based on the fact that adjacent frames in a short clip are highly coherent in content, we propose to extend the one-stage FiFo framework to a clip-in clip-out (CiCo) one, which performs VIS clip by clip. Specifically, we stack FPN features of all frames in a short video clip to build a spatio-temporal feature cube, and replace the 2D conv layers in the prediction heads and the mask branch with 3D conv layers, forming clip-level prediction heads (CPH) and clip-level mask heads (CMH). Then the clip-level masks of an instance can be generated by feeding its box-level predictions from CPH and clip-level features from CMH into a small fully convolutional network. A clip-level segmentation loss is proposed to ensure that the generated instance masks are temporally coherent in the clip. The proposed CiCo strategy is free of inter-frame alignment, and can be easily embedded into existing FiFo based VIS approaches. To validate the generality and effectiveness of our CiCo strategy, we apply it to two representative FiFo methods, Yolact \cite{bolya2019yolact} and CondInst \cite{tian2020conditional}, resulting in two new one-stage VIS models, namely CiCo-Yolact and CiCo-CondInst, which achieve 37.1/37.3\%, 35.2/35.4\% and 17.2/18.0\% mask AP using the ResNet50 backbone, and 41.8/41.4\%, 38.0/38.9\% and 18.0/18.2\% mask AP using the Swin Transformer tiny backbone on YouTube-VIS 2019, 2021 and OVIS valid sets, respectively, recording new state-of-the-arts. Code and video demos of CiCo can be found at \url{https://github.com/MinghanLi/CiCo}.
Blind image super-resolution (BISR) aims to reconstruct a high-resolution image from its low-resolution counterpart degraded by unknown blur kernel and noise. Many deep neural network based methods have been proposed to tackle this challenging problem without considering the image degradation model. However, they largely rely on the training sets and often fail to handle images with unseen blur kernels during inference. Deep unfolding methods have also been proposed to perform BISR by utilizing the degradation model. Nonetheless, the existing deep unfolding methods cannot explicitly solve the data term of the unfolding objective function, limiting their capability in blur kernel estimation. In this work, we propose a novel unfolded deep kernel estimation (UDKE) method, which, for the first time to our best knowledge, explicitly solves the data term with high efficiency. The UDKE based BISR method can jointly learn image and kernel priors in an end-to-end manner, and it can effectively exploit the information in both training data and image degradation model. Experiments on benchmark datasets and real-world data demonstrate that the proposed UDKE method could well predict complex unseen non-Gaussian blur kernels in inference, achieving significantly better BISR performance than state-of-the-art. The source code of UDKE is available at: https://github.com/natezhenghy/UDKE.