Score distillation has emerged as one of the most prevalent approaches for text-to-3D asset synthesis. Essentially, score distillation updates 3D parameters by lifting and back-propagating scores averaged over different views. In this paper, we reveal that the gradient estimation in score distillation is inherent to high variance. Through the lens of variance reduction, the effectiveness of SDS and VSD can be interpreted as applications of various control variates to the Monte Carlo estimator of the distilled score. Motivated by this rethinking and based on Stein’s identity, we propose a more general solution to reduce variance for score distillation, termed \emph{Stein Score Distillation (SSD)}. SSD incorporates control variates constructed by Stein identity, allowing for arbitrary baseline functions. This enables us to include flexible guidance priors and network architectures to explicitly optimize for variance reduction. In our experiments, the overall pipeline, dubbed \emph{SteinDreamer}, is implemented by instantiating the control variate with a monocular depth estimator. The results show that SSD can effectively reduce the distillation variance and consistently improve visual quality for both object- and scene-level generation.

Recent advancements in 2D and multimodal models have achieved remarkable success by leveraging large-scale training on extensive datasets. However, extending these achievements to enable free-form interactions and high-level semantic operations with complex 3D/4D scenes remains challenging. This difficulty stems from the limited availability of large-scale, annotated 3D/4D or multi-view datasets, which are crucial for generalizable vision and language tasks such as open-vocabulary and prompt-based segmentation, language-guided editing, and visual question answering (VQA). In this paper, we introduce Feature4X, a universal framework designed to extend any functionality from 2D vision foundation model into the 4D realm, using only monocular video input, which is widely available from user-generated content. The "X" in Feature4X represents its versatility, enabling any task through adaptable, model-conditioned 4D feature field distillation. At the core of our framework is a dynamic optimization strategy that unifies multiple model capabilities into a single representation. Additionally, to the best of our knowledge, Feature4X is the first method to distill and lift the features of video foundation models (e.g. SAM2, InternVideo2) into an explicit 4D feature field using Gaussian Splatting. Our experiments showcase novel view segment anything, geometric and appearance scene editing, and free-form VQA across all time steps, empowered by LLMs in feedback loops. These advancements broaden the scope of agentic AI applications by providing a foundation for scalable, contextually and spatiotemporally aware systems capable of immersive dynamic 4D scene interaction.

Rotated object detection aims to locate and classify objects with arbitrary orientations. In complex backgrounds, small rotated objects with limited salient features present challenges for standard backbones to extract high-quality, discriminative features. Additionally, traditional single-stage detection heads suffer from spatial prediction biases due to misalignment between classification and localization tasks. To tackle these challenges, We propose a novel network architecture, YOLO-KED, which incorporates the KAGFusion Module (KAGFM), EdgeFusion Module (EFM), and Dynamic Alignment and Rotated Detection Head (DARH) to enhance the feature extraction precision of the backbone and dynamically align the classification and localization tasks. We have also released a new dataset, termed as ICDM, specifically for rotated object detection, containing 3,214 images across 11 common categories of industrial components. Experimental results demonstrate that YOLOKED outperforms existing state-of-the-art methods across multiple detection metrics, especially in complex scenarios and small object detection tasks. ICDM is available at https://github.com/twodian/ICDM.

Using the latent diffusion model has proven effective in developing novel 3D generation techniques. To harness the latent diffusion model, a key challenge is designing a high-fidelity and efficient representation that links the latent space and the 3D space. In this paper, we introduce Atlas Gaussians, a novel representation for feed-forward native 3D generation. Atlas Gaussians represent a shape as the union of local patches, and each patch can decode 3D Gaussians. We parameterize a patch as a sequence of feature vectors and design a learnable function to decode 3D Gaussians from the feature vectors. In this process, we incorporate UV-based sampling, enabling the generation of a sufficiently large, and theoretically infinite, number of 3D Gaussian points. The large amount of 3D Gaussians enables the generation of high-quality details. Moreover, due to local awareness of the representation, the transformer-based decoding procedure operates on a patch level, ensuring efficiency. We train a variational autoencoder to learn the Atlas Gaussians representation, and then apply a latent diffusion model on its latent space for learning 3D Generation. Experiments show that our approach outperforms the prior arts of feed-forward native 3D generation. Project page: https://yanghtr.github.io/projects/atlas_gaussians.

The blooming of virtual reality and augmented reality (VR/AR) technologies has driven an increasing demand for the creation of high-quality, immersive, and dynamic environments. However, existing generative techniques either focus solely on dynamic objects or perform outpainting from a single perspective image, failing to meet the requirements of VR/AR applications that need free-viewpoint, 360$^{\circ}$ virtual views where users can move in all directions. In this work, we tackle the challenging task of elevating a single panorama to an immersive 4D experience. For the first time, we demonstrate the capability to generate omnidirectional dynamic scenes with 360$^{\circ}$ views at 4K (4096 $\times$ 2048) resolution, thereby providing an immersive user experience. Our method introduces a pipeline that facilitates natural scene animations and optimizes a set of 3D Gaussians using efficient splatting techniques for real-time exploration. To overcome the lack of scene-scale annotated 4D data and models, especially in panoramic formats, we propose a novel Panoramic Denoiser that adapts generic 2D diffusion priors to animate consistently in 360$^{\circ}$ images, transforming them into panoramic videos with dynamic scenes at targeted regions. Subsequently, we propose Dynamic Panoramic Lifting to elevate the panoramic video into a 4D immersive environment while preserving spatial and temporal consistency. By transferring prior knowledge from 2D models in the perspective domain to the panoramic domain and the 4D lifting with spatial appearance and geometry regularization, we achieve high-quality Panorama-to-4D generation at a resolution of 4K for the first time. Project page: https://4k4dgen.github.io/.

In recent years there have been remarkable breakthroughs in image-to-video generation. However, the 3D consistency and camera controllability of generated frames have remained unsolved. Recent studies have attempted to incorporate camera control into the generation process, but their results are often limited to simple trajectories or lack the ability to generate consistent videos from multiple distinct camera paths for the same scene. To address these limitations, we introduce Cavia, a novel framework for camera-controllable, multi-view video generation, capable of converting an input image into multiple spatiotemporally consistent videos. Our framework extends the spatial and temporal attention modules into view-integrated attention modules, improving both viewpoint and temporal consistency. This flexible design allows for joint training with diverse curated data sources, including scene-level static videos, object-level synthetic multi-view dynamic videos, and real-world monocular dynamic videos. To the best of our knowledge, Cavia is the first framework that enables users to generate multiple videos of the same scene with precise control over camera motion, while simultaneously preserving object motion. Extensive experiments demonstrate that Cavia surpasses state-of-the-art methods in terms of geometric consistency and perceptual quality.

Video generation is experiencing a rapid growth, driven by advances in diffusion models and the development of better and larger datasets. However, producing high-quality videos remains challenging, due to the high dimensional data and the complexity of the task. Recent efforts have primarily focused on enhancing visual quality and addressing temporal inconsistencies, such as flickering. Despite progress in these areas, the generated videos often fall short in terms of motion complexity and physical plausibility, with many outputs either appearing static or exhibiting unrealistic motion. In this work, we propose a framework aimed at improving the realism of motion in generated videos, exploring a complementary direction to much of the existing literature. Specifically, we advocate for the incorporation of a retrieval mechanism during the generation phase. The retrieved videos act as grounding signals, providing the model with demonstrations of how the objects move. Our pipeline is designed to apply to any text-to-video diffusion model, conditioning a pre-trained model on the retrieved samples with minimal fine-tuning. We demonstrate the superiority of our approach through established metrics, recently proposed benchmarks, and qualitative results, and we highlight additional applications of the framework. Code available at: https://github.com/helia95/ragme.

With the rapid advancement and widespread adoption of VR/AR technologies, there is a growing demand for the creation of high-quality, immersive dynamic scenes. However, existing generation works predominantly concentrate on the creation of static scenes or narrow perspective-view dynamic scenes, falling short of delivering a truly 360-degree immersive experience from any viewpoint. In this paper, we introduce TiP4GEN, an advanced text-to-dynamic panorama scene generation framework that enables fine-grained content control and synthesizes motion-rich, geometry-consistent panoramic 4D scenes. TiP4GEN integrates panorama video generation and dynamic scene reconstruction to create 360-degree immersive virtual environments. For video generation, we introduce a Dual-branch Generation Model consisting of a panorama branch and a perspective branch, responsible for global and local view generation, respectively. A bidirectional cross-attention mechanism facilitates comprehensive information exchange between the branches. For scene reconstruction, we propose a Geometry-aligned Reconstruction Model based on 3D Gaussian Splatting. By aligning spatial-temporal point clouds using metric depth maps and initializing scene cameras with estimated poses, our method ensures geometric consistency and temporal coherence for the reconstructed scenes. Extensive experiments demonstrate the effectiveness of our proposed designs and the superiority of TiP4GEN in generating visually compelling and motion-coherent dynamic panoramic scenes.

Text-to-image generative models are becoming increasingly popular and accessible to the general public. As these models see large-scale deployments, it is necessary to deeply investigate their safety and fairness to not disseminate and perpetuate any kind of biases. However, existing works focus on detecting closed sets of biases defined a priori, limiting the studies to well-known concepts. In this paper, we tackle the challenge of open-set bias detection in text-to-image generative models presenting OpenBias, a new pipeline that identifies and quantifies the severity of biases agnostically, without access to any precompiled set. OpenBias has three stages. In the first phase, we leverage a Large Language Model (LLM) to propose biases given a set of captions. Secondly, the target generative model produces images using the same set of captions. Lastly, a Vision Question Answering model recognizes the presence and extent of the previously proposed biases. We study the behavior of Stable Diffusion 1.5, 2, and XL emphasizing new biases, never investigated before. Via quantitative experiments, we demonstrate that OpenBias agrees with current closed-set bias detection methods and human judgement.

Generative image editing has recently witnessed extremely fast-paced growth. Some works use high-level conditioning such as text, while others use low-level conditioning. Nevertheless, most of them lack fine-grained control over the properties of the different objects present in the image, i.e. object-level image editing. In this work, we tackle the task by perceiving the images as an amalgamation of various objects and aim to control the properties of each object in a fine-grained manner. Out of these properties, we identify structure and appearance as the most intuitive to understand and useful for editing purposes. We propose PAIR Diffusion, a generic framework that enables a diffusion model to control the structure and appearance properties of each object in the image. We show that having control over the properties of each object in an image leads to comprehensive editing capabilities. Our framework allows for various object-level editing operations on real images such as reference image-based appearance editing, free-form shape editing, adding objects, and variations. Thanks to our design, we do not require any inversion step. Additionally, we propose multi-modal classifier-free guidance which enables editing images using both reference images and text when using our approach with foundational diffusion models. We validate the above claims by extensively evaluating our framework on both unconditional and foundational diffusion models.

Despite the remarkable performance of score distillation in text-to-3D generation, such techniques notoriously suf-fer from view inconsistency issues, also known as “Janus” artifact, where the generated objects fake each view with multiple front faces. Although empirically effective methods have approached this problem via score debiasing or prompt engineering, a more rigorous perspective to explain and tackle this problem remains elusive. In this paper, we reveal that the existing score distillation-based text-to-3D generation frameworks degenerate to maximal likelihood seeking on each view independently and thus suffer from the mode collapse problem, manifesting as the Janus artifact in practice. To tame mode collapse, we improve score distillation by re-establishing the entropy term in the corresponding variational objective, which is applied to the distribution of rendered images. Maximizing the entropy encourages diversity among different views in generated 3D assets, thereby mitigating the Janus problem. Based on this new objective, we derive a new update rule for 3D score distillation, dubbed Entropic Score Distillation (ESD). We theoretically reveal that ESD can be simplified and implemented by just adopting the classifier-free guidance trick upon variational score distillation. Although embarrassingly straightforward, our extensive experiments demonstrate that ESD can be an effective treatment for Janus artifacts in score distillation.

3D scene representations have gained immense popularity in recent years. Methods that use Neural Radiance fields are versatile for traditional tasks such as novel view synthesis. In recent times, some work has emerged that aims to extend the functionality of NeRF beyond view synthesis, for semantically aware tasks such as editing and segmentation using 3D feature field distillation from 2D foundation models. However, these methods have two major limitations: (a) they are limited by the rendering speed of NeRF pipelines, and (b) implicitly represented feature fields suffer from continuity artifacts reducing feature quality. Recently, 3D Gaussian Splatting has shown state-of-the-art performance on real-time radiance field rendering. In this work, we go one step further: in addition to radiance field rendering, we enable 3D Gaussian splatting on arbitrary-dimension semantic features via 2D foundation model distillation. This translation is not straightforward: naively incorporating feature fields in the 3DGS framework encounters significant challenges, notably the disparities in spatial resolution and channel consistency between RGB images and feature maps. We propose architectural and training changes to efficiently avert this problem. Our proposed method is general, and our experiments showcase novel view semantic segmentation, language-guided editing and segment anything through learning feature fields from state-of-the-art 2D foundation models such as SAM and CLIP-LSeg. Across experiments, our distillation method is able to provide comparable or better results, while being significantly faster to both train and render. Additionally, to the best of our knowledge, we are the first method to enable point and bounding-box prompting for radiance field manipulation, by leveraging the SAM model. Project website at: https://feature-3dgs.github.io/.

Recent advances in real-time neural rendering using point-based techniques have enabled broader adoption of 3D representations. However, foundational approaches like 3D Gaussian Splatting impose substantial storage overhead, as Structure-from-Motion (SfM) points can grow to millions, often requiring gigabyte-level disk space for a single unbounded scene. This growth presents scalability challenges and hinders splatting efficiency. To address this, we introduce LightGaussian, a method for transforming 3D Gaussians into a more compact format. Inspired by Network Pruning, LightGaussian identifies Gaussians with minimal global significance on scene reconstruction, and applies a pruning and recovery process to reduce redundancy while preserving visual quality. Knowledge distillation and pseudo-view augmentation then transfer spherical harmonic coefficients to a lower degree, yielding compact representations. Gaussian Vector Quantization, based on each Gaussian’s global significance, further lowers bitwidth with minimal accuracy loss. LightGaussian achieves an average 15 times compression rate while boosting FPS from 144 to 237 within the 3D-GS framework, enabling efficient complex scene representation on the Mip-NeRF 360 and Tank & Temple datasets. The proposed Gaussian pruning approach is also adaptable to other 3D representations (e.g., Scaffold-GS), demonstrating strong generalization capabilities.

The availability of large-scale multimodal datasets and advancements in diffusion models have significantly accelerated progress in 4D content generation. Most prior approaches rely on multiple images or video diffusion models, utilizing score distillation sampling for optimization or generating pseudo novel views for direct supervision. However, these methods are hindered by slow optimization speeds and multi-view inconsistency issues. Spatial and temporal consistency in 4D geometry has been extensively explored respectively in 3D-aware diffusion models and traditional monocular video diffusion models. Building on this foundation, we propose a strategy to migrate the temporal consistency in video diffusion models to the spatial-temporal consistency required for 4D generation. Specifically, we present a novel framework, \textbf{Diffusion4D}, for efficient and scalable 4D content generation. Leveraging a meticulously curated dynamic 3D dataset, we develop a 4D-aware video diffusion model capable of synthesizing orbital views of dynamic 3D assets. To control the dynamic strength of these assets, we introduce a 3D-to-4D motion magnitude metric as guidance. Additionally, we propose a novel motion magnitude reconstruction loss and 3D-aware classifier-free guidance to refine the learning and generation of motion dynamics. After obtaining orbital views of the 4D asset, we perform explicit 4D construction with Gaussian splatting in a coarse-to-fine manner. Extensive experiments demonstrate that our method surpasses prior state-of-the-art techniques in terms of generation efficiency and 4D geometry consistency across various prompt modalities.

Neural Radiance Fields (NeRFs) are a powerful representation for modeling a 3D scene as a continuous function. Though NeRF is able to render complex 3D scenes with view-dependent effects, few efforts have been devoted to exploring its limits in a high-resolution setting. Specifically, existing NeRF-based methods face several limitations when reconstructing high-resolution real scenes, including a very large number of parameters, misaligned input data, and overly smooth details. In this work, we conduct the first pilot study on training NeRF with high-resolution data and propose the corresponding solutions: 1) marrying the multilayer perceptron (MLP) with convolutional layers which can encode more neighborhood information while reducing the total number of parameters; 2) a novel training strategy to address misalignment caused by moving objects or small camera calibration errors; and 3) a high-frequency aware loss. Our approach is nearly free without introducing obvious training/testing costs, while experiments on different datasets demonstrate that it can recover more high-frequency details compared with the current state-of-the-art NeRF models. Project page: https://yifanjiang19.github.io/alignerf.

Virtual reality and augmented reality (XR) bring increasing demand for 3D content generation. However, creating high-quality 3D content requires tedious work from a human expert. In this work, we study the challenging task of lifting a single image to a 3D object and, for the first time, demonstrate the ability to generate a plausible 3D object with 360° views that corresponds well with the given reference image. By conditioning on the reference image, our model can fulfill the everlasting curiosity for synthesizing novel views of objects from images. Our technique sheds light on a promising direction of easing the workflows for 3D artists and XR designers. We propose a novel framework, dubbed NeuralLift-360, that utilizes a depth-aware neural radiance representation (NeRF) and learns to craft the scene guided by denoising diffusion models. By introducing a ranking loss, our NeuralLift-360 can be guided with rough depth estimation in the wild. We also adopt a CLIP-guided sampling strategy for the diffusion prior to provide coherent guidance. Extensive experiments demonstrate that our NeuralLift-360 significantly outperforms existing state-of-the-art baselines. Project page: https://vita-group.github.io/NeuralLift-360/

An increasing number of researchers are finding use for nth-order gradient computations for a wide variety of applications, including graphics, meta-learning (MAML), scientific computing, and most recently, implicit neural representations (INRs). Recent work shows that the gradient of an INR can be used to edit the data it represents directly without needing to convert it back to a discrete representation. However, given a function represented as a computation graph, traditional architectures face challenges in efficiently computing its nth-order gradient due to the higher demand for computing power and higher complexity in data movement. This makes it a promising target for FPGA acceleration. In this work, we introduce INR-Arch, a framework that transforms the computation graph of an nth-order gradient into a hardware-optimized dataflow architecture. We address this problem in two phases. First, we design a dataflow architecture that uses FIFO streams and an optimized computation kernel library, ensuring high memory efficiency and parallel computation. Second, we propose a compiler that extracts and optimizes computation graphs, automatically configures hardware parameters such as latency and stream depths to optimize throughput, while ensuring deadlock-free operation, and outputs High-Level Synthesis (HLS) code for FPGA implementation. We utilize INR editing as our benchmark, presenting results that demonstrate 1.8-4.8x and 1.5-3.6x speedup compared to CPU and GPU baselines respectively. Furthermore, we obtain 3.1-8.9x and 1.7-4.3x lower memory usage, and 1.7-11.3x and 5.5-32.8x lower energy-delay product. Our framework will be made open-source and available on GitHub.****https://github.com/sharc-lab/inr-arch

For a complicated algorithm, its implementation by a human programmer usually starts with outlining a rough control flow followed by iterative enrichments, eventually yielding carefully generated syntactic structures and variables in a hierarchy. However, state-of-the-art large language models generate codes in a single pass, without intermediate warm-ups to reflect the structured thought process of "outline-then-detail". Inspired by the recent success of chain-of-thought prompting, we propose ChainCoder, a program synthesis language model that generates Python code progressively, i.e. from coarse to fine in multiple passes. We first decompose source code into layout frame components and accessory components via abstract syntax tree parsing to construct a hierarchical representation. We then reform our prediction target into a multi-pass objective, each pass generates a subsequence, which is concatenated in the hierarchy. Finally, a tailored transformer architecture is leveraged to jointly encode the natural language descriptions and syntactically aligned I/O data samples. Extensive evaluations show that ChainCoder outperforms state-of-the-arts, demonstrating that our progressive generation eases the reasoning procedure and guides the language model to generate higher-quality solutions. Our codes are available at: https://github.com/VITA-Group/ChainCoder.

Low light enhancement has gained increasing importance with the rapid development of visual creation and editing. However, most existing enhancement algorithms are designed to homogeneously increase the brightness of images to a pre-defined extent, limiting the user experience. To address this issue, we propose Controllable Light Enhancement Diffusion Model, dubbed CLE Diffusion, a novel diffusion framework to provide users with rich controllability.Built with a conditional diffusion model, we introduce an illumination embedding to let users control their desired brightness level. Additionally, we incorporate the Segment-Anything Model (SAM) to enable user-friendly region controllability, where users can click on objects to specify the regions they wish to enhance. Extensive experiments demonstrate that CLE Diffusion achieves competitive performance regarding quantitative metrics, qualitative results, and versatile controllability. Project page: https://yuyangyin.github.io/CLEDiffusion

Colored glass, which is commonly seen in modern city life, often degrades images taken through it with co-occurring reflection and color bias due to its optical property of simultaneous transmission, reflection, and wavelength-selective absorption. Recovering the clean background behind colored glass is inherently challenging due to the mutual interference of two degradations within a single mixture observation, and has barely been specifically considered by existing image restoration methods. In this paper, we aim at realizing faithful background scene recovery for an image taken in front of colored glass. We first analyze the formation model of mixed degradations caused by colored glass, and propose a cooperative framework to address the mutual interference problem, featuring a novel glass color invariant loss and progressive refinement. Besides, we propose a data synthesis strategy for network training. Experimental results on our newly collected real-world dataset show that our proposed method achieves state-of-the-art performance.

Have you ever looked up at the sky and imagined what the clouds look like? In this work, we present an interesting task that augments clouds in the sky with imagined sketches. Different from generic image-to-sketch translation tasks, unique challenges are introduced: real-world clouds have different levels of similarity to something; sketch generation without sketch retrieval could lead to something unrecognizable; a retrieved sketch from some dataset cannot be directly used because of the mismatch of the shape; an optimal sketch imagination is subjective. We propose Cloud2Sketch, a novel self-supervised pipeline to tackle the aforementioned challenges. First, we pre-process cloud images with a cloud detector and a thresholding algorithm to obtain cloud contours. Then, cloud contours are passed through a retrieval module to retrieve sketches with similar geometrical shapes. Finally, we adopt a novel sketch translation model with built-in free-form deformation for aligning the sketches to cloud contours. To facilitate training, an icon-based sketch collection named Sketchy Zoo is proposed. Extensive experiments validate the effectiveness of our method both qualitatively and quantitatively.

Low-light enhancement is an increasingly important function in image editing and visual creation. Most existing enhancing algorithms are trained to enlighten a given image in a globally homogeneous way, and (implicitly) to some predefined extent of brightness. They are neither capable of enhancing only local regions of interest ("where") while keeping the overall visual appearance plausible, nor producing outputs at a range of different illumination levels ("how much"). Those hurdles significantly limit the prospect of flexible, customizable, or even user-interactive low-light enhancement. To address these gaps, we propose Region-Controllable Robust Light Enhancement (ReCoRo), a novel framework that allows users to directly specify "where" and "how much" they want to enhance from an input low-light image; meanwhile, the model will learn to intelligently maintain the overall consistent visual appearance and plausible composition via a discriminator. Moreover, since in practical mobile APPs, such user specifications often come in imprecise forms (e.g., finger-drawn masks), we propose to bake in domain-specific data augmentations into training ReCoRo, so that the learned model can gain resilience to various roughly-supplied user masks. Up to our best knowledge, ReCoRo is the first of its kind that allows the user to localize the enlightenment region as well as to control the light intensity. Extensive experiments clearly demonstrate that ReCoRo outperforms state-of-the-art methods in terms of qualitative results, quantitative metrics, and versatile controllability. Project repository: https://bit.ly/ReCoRo-lowlight.

Implicit Neural Representations (INRs) encoding continuous multi-media data via multi-layer perceptrons has shown undebatable promise in various computer vision tasks. Despite many successful applications, editing and processing an INR remains intractable as signals are represented by latent parameters of a neural network. Existing works manipulate such continuous representations via processing on their discretized instance, which breaks down the compactness and continuous nature of INR. In this work, we present a pilot study on the question: how to directly modify an INR without explicit decoding? We answer this question by proposing an implicit neural signal processing network, dubbed INSP-Net, via differential operators on INR. Our key insight is that spatial gradients of neural networks can be computed analytically and are invariant to translation, while mathematically we show that any continuous convolution filter can be uniformly approximated by a linear combination of high-order differential operators. With these two knobs, INSP-Net instantiates the signal processing operator as a weighted composition of computational graphs corresponding to the high-order derivatives of INRs, where the weighting parameters can be data-driven learned. Based on our proposed INSP-Net, we further build the first Convolutional Neural Network (CNN) that implicitly runs on INRs, named INSP-ConvNet. Our experiments validate the expressiveness of INSP-Net and INSP-ConvNet in fitting low-level image and geometry processing kernels (e.g. blurring, deblurring, denoising, inpainting, and smoothening) as well as for high-level tasks on implicit fields such as image classification.

Data-driven rain streak removal methods, which most of rely on synthesized paired data, usually come across the generalization problem when being applied in real cases. In this paper, we propose a novel deep-learning based rain streak removal method injected with self-supervision to improve the ability to remove rain streaks in various scales. To realize this goal, we made efforts in two aspects. First, considering that rain streak removal is highly correlated with texture characteristics, we create a fractal band learning (FBL) network based on frequency band recovery. It integrates commonly seen band feature operations with neural modules and effectively improves the capacity to capture discriminative features for deraining. Second, to further improve the generalization ability of FBL for rain streaks in various scales, we add cross-scale self-supervision to regularize the network training. The constraint forces the extracted features of inputs in different scales to be equivalent after rescaling. Therefore, FBL can offer similar responses based on solely image content without the interleave of scale and is capable to remove rain streaks in various scales. Extensive experiments in quantitative and qualitative evaluations demonstrate the superiority of our FBL for rain streak removal, especially for the real cases where very large rain streaks exist, and prove the effectiveness of its each component. Our code will be public available at: https://github.com/flyywh/AAAI-2020-FBL-SS.