We propose a generative model that, given a coarsely edited image, synthesizes a photorealistic output that follows the prescribed layout. Our method transfers fine details from the original image and preserve the identity of its parts. Yet, it adapts it to the lighting and context defined by the new layout. Our key insight is that videos are a powerful source of supervision for this task: objects and camera motions provide many observations of how the world changes with viewpoint, lighting, and physical interactions. We construct an image dataset in which each sample is a pair of source and target frames extracted from the same video at randomly chosen time intervals. We warp the source frame toward the target using two motion models that mimic the expected test-time user edits. We supervise our model to translate the warped image into the ground truth, starting from a pretrained diffusion model. Our model design explicitly enables fine detail transfer from the source frame to the generated image, while closely following the user-specified layout. We show that by using simple segmentations and coarse 2D manipulations, we can synthesize a photorealistic edit faithful to the user’s input while addressing second-order effects like harmonizing the lighting and physical interactions between edited objects. Project page and code can be found at https://magic-fixup.github.io.

We present a novel method to differentiate integrals of discontinuous functions, which are common in inverse graphics, computer vision, and machine learning applications. Previous methods either require specialized routines to sample the discontinuous boundaries of predetermined primitives, or use reparameterization techniques that suffer from high variance. In contrast, our method handles general discontinuous functions, expressed as shader programs, without requiring manually specified boundary sampling routines. We achieve this through a program transformation that converts discontinuous functions into piecewise constant ones, enabling efficient boundary sampling through a novel segment snapping technique, and accurate derivatives at the boundary by simply comparing values on both sides of the discontinuity. Our method handles both explicit boundaries (polygons, ellipses, Bézier curves) and implicit ones (neural networks, noise-based functions, swept surfaces). We demonstrate that our system supports a wide range of applications, including painterly rendering, raster image fitting, constructive solid geometry, swept surfaces, mosaicing, and ray marching.

How to frame (or crop) a photo often depends on the image subject and its context; e.g., a human portrait. Recent works have defined the subject-aware image cropping task as a nuanced and practical version of image cropping. We propose a weakly-supervised approach (GenCrop) to learn what makes a high-quality, subject-aware crop from professional stock images. Unlike supervised prior work, GenCrop requires no new manual annotations beyond the existing stock image collection. The key challenge in learning from this data, however, is that the images are already cropped and we do not know what regions were removed. Our insight is to combine a library of stock images with a modern, pre-trained text-to-image diffusion model. The stock image collection provides diversity, and its images serve as pseudo-labels for a good crop. The text-image diffusion model is used to out-paint (i.e., outward inpainting) realistic uncropped images. Using this procedure, we are able to automatically generate a large dataset of cropped-uncropped training pairs to train a cropping model. Despite being weakly-supervised, GenCrop is competitive with state-of-the-art supervised methods and significantly better than comparable weakly-supervised baselines on quantitative and qualitative evaluation metrics.

Diffusion models have shown an impressive ability to model complex data distributions, with several key advantages over GANs, such as stable training, better coverage of the training distribution's modes, and the ability to solve inverse problems without extra training. However, most diffusion models learn the distribution of fixed-resolution images. We propose to learn the distribution of continuous images by training diffusion models on image neural fields, which can be rendered at any resolution, and show its advantages over fixed-resolution models. To achieve this, a key challenge is to obtain a latent space that represents photorealistic image neural fields. We propose a simple and effective method, inspired by several recent techniques but with key changes to make the image neural fields photo-realistic. Our method can be used to convert existing latent diffusion autoencoders into image neural field autoen-coders. We show that image neural field diffusion models can be trained using mixed-resolution image datasets, outperform fixed-resolution diffusion models followed by super-resolution models, and can solve inverse problems with conditions applied at different scales efficiently.

We present VecFusion, a new neural architecture that can generate vector fonts with varying topological structures and precise control point positions. Our approach is a cascaded diffusion model which consists of a raster diffusion model followed by a vector diffusion model. The raster model generates low-resolution, rasterized fonts with auxiliary control point information, capturing the global style and shape of the font, while the vector model synthesizes vector fonts conditioned on the low-resolution raster fonts from the first stage. To synthesize long and complex curves, our vector diffusion model uses a transformer architecture and a novel vector representation that enables the modeling of diverse vector geometry and the precise prediction of control points. Our experiments show that, in contrast to previous generative models for vector graphics, our new cascaded vector diffusion model generates higher quality vector fonts, with complex structures and diverse styles.

Diffusion models generate high-quality images but require dozens of forward passes. We introduce Distribution Matching Distillation (DMD), a procedure to transform a diffusion model into a one-step image generator with minimal impact on image quality. We enforce the one-step image generator match the diffusion model at distribution level, by minimizing an approximate KL divergence whose gradient can be expressed as the difference between 2 score functions, one of the target distribution and the other of the synthetic distribution being produced by our one-step generator. The score functions are parameterized as two diffusion models trained separately on each distribution. Combined with a simple regression loss matching the large-scale structure of the multi-step diffusion outputs, our method outperforms all published few-step diffusion approaches, reaching 2.62 FID on ImageNet 64×64 and 11.49 FID on zero-shot COCO-30k, comparable to Stable Diffusion but orders of magnitude faster. Utilizing FP16 inference, our model can generate images at 20 FPS on modern hardware.

Recent approaches have shown promises distilling expensive diffusion models into efficient one-step generators. Amongst them, Distribution Matching Distillation (DMD) produces one-step generators that match their teacher in distribution, i.e., the distillation process does not enforce a one-to-one correspondence with the sampling trajectories of their teachers. However, to ensure stable training in practice, DMD requires an additional regression loss computed using a large set of noise--image pairs, generated by the teacher with many steps of a deterministic sampler. This is not only computationally expensive for large-scale text-to-image synthesis, but it also limits the student's quality, tying it too closely to the teacher's original sampling paths. We introduce DMD2, a set of techniques that lift this limitation and improve DMD training. First, we eliminate the regression loss and the need for expensive dataset construction. We show that the resulting instability is due to the "fake" critic not estimating the distribution of generated samples with sufficient accuracy and propose a two time-scale update rule as a remedy. Second, we integrate a GAN loss into the distillation procedure, discriminating between generated samples and real images. This lets us train the student model on real data, thus mitigating the imperfect "real" score estimation from the teacher model, and thereby enhancing quality. Third, we introduce a new training procedure that enables multi-step sampling in the student, and addresses the training--inference input mismatch of previous work, by simulating inference-time generator samples during training. Taken together, our improvements set new benchmarks in one-step image generation, with FID scores of 1.28 on ImageNet-64×64 and 8.35 on zero-shot COCO 2014, surpassing the original teacher despite a 500X reduction in inference cost. Further, we show our approach can generate megapixel images by distilling SDXL, demonstrating exceptional visual quality among few-step methods, and surpassing the teacher. We release our code and pretrained models.

Can one inject new concepts into an already trained generative model, while respecting its existing structure and knowledge? We propose a new task - domain expansion - to address this. Given a pretrained generator and novel (but related) domains, we expand the generator to jointly model all domains, old and new, harmoniously. First, we note the generator contains a meaningful, pretrained latent space. Is it possible to minimally perturb this hard-earned representation, while maximally representing the new domains? Interestingly, we find that the latent space offers unused, “dormant” directions, which do not affect the output. This provides an opportunity: By “repurposing” these directions, we can represent new domains without perturbing the original representation. In fact, we find that pretrained generators have the capacity to add several- even hundreds - of new domains! Using our expansion method, one “expanded” model can supersede numerous domain-specific models, without expanding the model size. Additionally, a single expanded generator natively supports smooth transitions between domains, as well as composition of domains. Code and project page available here.

Learning-based image harmonization techniques are usually trained to undo synthetic random global transformations applied to a masked foreground in a single ground truth photo. This simulated data does not model many of the important appearance mismatches (illumination, object boundaries, etc.) between foreground and background in real composites, leading to models that do not generalize well and cannot model complex local changes. We propose a new semi-supervised training strategy that addresses this problem and lets us learn complex local appearance harmonization from unpaired real composites, where foreground and background come from different images. Our model is fully parametric. It uses RGB curves to correct the global colors and tone and a shading map to model local variations. Our method outperforms previous work on established benchmarks and real composites, as shown in a user study, and processes high-resolution images interactively. Code, and project page available at: https://kewang0622.github.io/sprih/

We introduce a self-supervised training strategy for burst super-resolution that only uses noisy low-resolution bursts during training. Our approach eliminates the need to carefully tune synthetic data simulation pipelines, which often do not match real-world image statistics. Compared to weakly-paired training strategies, which require noisy smartphone burst photos of static scenes, paired with a clean reference obtained from a tripod-mounted DSLR camera, our approach is more scalable, and avoids the color mismatch between the smartphone and DSLR. To achieve this, we propose a new self-supervised objective that uses a forward imaging model to recover a high-resolution image from aliased high frequencies in the burst. Our approach does not require any manual tuning of the forward model's parameters; we learn them from data. Furthermore, we show our training strategy is robust to dynamic scene motion in the burst, which enables training burst super-resolution models using in-the-wild data. Extensive experiments on real and synthetic data show that, despite only using noisy bursts during training, models trained with our self-supervised strategy match, and sometimes surpass, the quality of fully-supervised baselines trained with synthetic data or weakly-paired ground-truth. Finally, we show our training strategy is general using four different burst super-resolution architectures.

Compositing virtual objects into real background images requires one to carefully match the scene’s camera parameters, surface geometry, textures, and lighting to obtain plausible renderings. Recent learning approaches have shown many scene properties can be estimated from images, resulting in robust automatic single-image compositing systems, but many challenges remain. In particular, interactions between real and synthetic shadows are not handled gracefully by existing methods, which typically assume a shadow-free background. As a result, they tend to generate double shadows when the synthetic object’s cast shadow overlaps a background shadow, and ignore shadows from the background that should be cast onto the synthetic object. In this paper, we present a compositing method for outdoor scenes that addresses these issues and produces realistic cast shadows. This requires identifying existing shadows, including soft shadow boundaries, then reasoning about the ambiguity of unknown ground albedo and scene lighting to match the color and intensity of shaded areas. Using supervision from shadow removal and detection datasets, we propose a generative adversarial pipeline and improved composition equations that simultaneously handle both shadow interaction scenarios. We evaluate our method on challenging, real outdoor images from multiple distributions and datasets. Quantitative and qualitative comparisons show our approach produces more realistic results than existing alternatives. Our code, datasets, and trained models are publicly available at https://lvsn.github.io/shadowcompositing.

Neural image representations offer the possibility of high fidelity, compact storage, and resolution-independent accuracy, providing an attractive alternative to traditional pixel- and grid-based representations. However, coordinate neural networks fail to capture discontinuities present in the image and tend to blur across them; we aim to address this challenge. In many cases, such as rendered images, vector graphics, diffusion curves, or solutions to partial differential equations, the locations of the discontinuities are known. We take those locations as input, represented as linear, quadratic, or cubic Bézier curves, and construct a feature field that is discontinuous across these locations and smooth everywhere else. Finally, we use a shallow multi-layer perceptron to decode the features into the signal value. To construct the feature field, we develop a new data structure based on a curved triangular mesh, with features stored on the vertices and on a subset of the edges that are marked as discontinuous. We show that our method can be used to compress a 100, 000 2 -pixel rendered image into a 25MB file; can be used as a new diffusion-curve solver by combining with Monte-Carlo-based methods or directly supervised by the diffusion-curve energy; or can be used for compressing 2D physics simulation data.

Separating an image into meaningful underlying components is a crucial first step for both editing and understanding images. We present a method capable of selecting the regions of a photograph exhibiting the same material as an artist-chosen area. Our proposed approach is robust to shading, specular highlights, and cast shadows, enabling selection in real images. As we do not rely on semantic segmentation (different woods or metal should not be selected together), we formulate the problem as a similarity-based grouping problem based on a user-provided image location. In particular, we propose to leverage the unsupervised DINO [Caron et al. 2021] features coupled with a proposed Cross-Similarity Feature Weighting module and an MLP head to extract material similarities in an image. We train our model on a new synthetic image dataset, that we release. We show that our method generalizes well to real-world images. We carefully analyze our model's behavior on varying material properties and lighting. Additionally, we evaluate it against a hand-annotated benchmark of 50 real photographs. We further demonstrate our model on a set of applications, including material editing, in-video selection, and retrieval of object photographs with similar materials.

We present a method to automatically compute correct gradients with respect to geometric scene parameters in neural SDF renderers. Recent physically-based differentiable rendering techniques for meshes have used edge-sampling to handle discontinuities, particularly at object silhouettes, but SDFs do not have a simple parametric form amenable to sampling. Instead, our approach builds on area-sampling techniques and develops a continuous warping function for SDFs to account for these discontinuities. Our method leverages the distance to surface encoded in an SDF and uses quadrature on sphere tracer points to compute this warping function. We further show that this can be done by subsampling the points to make the method tractable for neural SDFs. Our differentiable renderer can be used to optimize neural shapes from multi-view images and produces comparable 3D reconstructions to recent SDF-based inverse rendering methods, without the need for 2D segmentation masks to guide the geometry optimization and no volumetric approximations to the geometry.

We present a method to automatically synthesize efficient, high-quality demosaicking algorithms, across a range of computational budgets, given a loss function and training data. It performs a multi-objective, discrete-continuous optimization which simultaneously solves for the program structure and parameters that best tradeoff computational cost and image quality. We design the method to exploit domain-specific structure for search efficiency. We apply it to several tasks, including demosaicking both Bayer and Fuji X-Trans color filter patterns, as well as joint demosaicking and super-resolution. In a few days on 8 GPUs, it produces a family of algorithms that significantly improves image quality relative to the prior state-of-the-art across a range of computational budgets from 10 s to 1000 s of operations per pixel (1 dB–3 dB higher quality at the same cost, or 8.5–200× higher throughput at same or better quality). The resulting programs combine features of both classical and deep learning-based demosaicking algorithms into more efficient hybrid combinations, which are bandwidth-efficient and vectorizable by construction. Finally, our method automatically schedules and compiles all generated programs into optimized SIMD code for modern processors.

Vector graphics are widely used to represent fonts, logos, digital artworks, and graphic designs. But, while a vast body of work has focused on generative algorithms for raster images, only a handful of options exists for vector graphics. One can always rasterize the input graphic and resort to image-based generative approaches, but this negates the advantages of the vector representation. The current alternative is to use specialized models that require explicit supervision on the vector graphics representation at training time. This is not ideal because large-scale high-quality vector-graphics datasets are difficult to obtain. Furthermore, the vector representation for a given design is not unique, so models that supervise on the vector representation are unnecessarily constrained. Instead, we propose a new neural network that can generate complex vector graphics with varying topologies, and only requires indirect supervision from readily-available raster training images (i.e., with no vector counterparts). To enable this, we use a differentiable rasterization pipeline that renders the generated vector shapes and composites them together onto a raster canvas. We demonstrate our method on a range of datasets, and provide comparison with state-of-the-art SVG-VAE and DeepSVG, both of which require explicit vector graphics supervision. Finally, we also demonstrate our approach on the MNIST dataset, for which no groundtruth vector representation is available. Source code, datasets and more results are available at http://geometry.cs.ucl.ac.uk/projects/2021/Im2Vec/.

We introduce a new generator architecture, aimed at fast and efficient high-resolution image-to-image translation. We design the generator to be an extremely lightweight function of the full-resolution image. In fact, we use pixel-wise networks; that is, each pixel is processed independently of others, through a composition of simple affine transformations and nonlinearities. We take three important steps to equip such a seemingly simple function with adequate expressivity. First, the parameters of the pixel-wise networks are spatially varying, so they can represent a broader function class than simple 1 × 1 convolutions. Second, these parameters are predicted by a fast convolutional network that processes an aggressively low-resolution representation of the input. Third, we augment the input image by concatenating a sinusoidal encoding of spatial coordinates, which provides an effective inductive bias for generating realistic novel high-frequency image content. As a result, our model is up to 18× faster than state-of-the-art baselines. We achieve this speedup while generating comparable visual quality across different image resolutions and translation domains.

We introduce a neural network-based method to denoise pairs of images taken in quick succession, with and without a flash, in low-light environments. Our goal is to produce a high-quality rendering of the scene that preserves the color and mood from the ambient illumination of the noisy no-flash image, while recovering surface texture and detail revealed by the flash. Our network outputs a gain map and a field of kernels, the latter obtained by linearly mixing elements of a per-image low-rank kernel basis. We first apply the kernel field to the no-flash image, and then multiply the result with the gain map to create the final output. We show our network effectively learns to produce high-quality images by combining a smoothed out estimate of the scene’s ambient appearance from the no-flash image, with high-frequency albedo details extracted from the flash input. Our experiments show significant improvements over alternative captures without a flash, and baseline denoisers that use flash no-flash pairs. In particular, our method produces images that are both noise-free and contain accurate ambient colors without the sharp shadows or strong specular highlights visible in the flash image.

Human pose is a useful feature for fine-grained sports action understanding. However, pose estimators are often unreliable when run on sports video due to domain shift and factors such as motion blur and occlusions. This leads to poor accuracy when downstream tasks, such as action recognition, depend on pose. End-to-end learning circumvents pose, but requires more labels to generalize.We introduce Video Pose Distillation (VPD), a weakly-supervised technique to learn features for new video domains, such as individual sports that challenge pose estimation. Under VPD, a student network learns to extract robust pose features from RGB frames in the sports video, such that, whenever pose is considered reliable, the features match the output of a pretrained teacher pose detector. Our strategy retains the best of both pose and end-to-end worlds, exploiting the rich visual patterns in raw video frames, while learning features that agree with the athletes’ pose and motion in the target video domain to avoid over-fitting to patterns unrelated to athletes’ motion.VPD features improve performance on few-shot, fine-grained action recognition, retrieval, and detection tasks in four real-world sports video datasets, without requiring additional ground-truth pose annotations.

Multi-Layer Perceptrons (MLPs) make powerful functional representations for sampling and reconstruction problems involving low-dimensional signals like images, shapes and light fields. Recent works have significantly improved their ability to represent high-frequency content by using periodic activations or positional encodings. This often came at the expense of generalization: modern methods are typically optimized for a single signal. We present a new representation that generalizes to multiple instances and achieves state-of-the-art fidelity. We use a dual-MLP architecture to encode the signals. A synthesis network creates a functional mapping from a low-dimensional input (e.g. pixel-position) to the output domain (e.g. RGB color). A modulation network maps a latent code corresponding to the target signal to parameters that modulate the periodic activations of the synthesis network. We also propose a local-functional representation which enables generalization. The signal’s domain is partitioned into a regular grid, with each tile represented by a latent code. At test time, the signal is encoded with high-fidelity by inferring (or directly optimizing) the latent code-book. Our approach produces generalizable functional representations of images, videos and shapes, and achieves higher reconstruction quality than prior works that are optimized for a single signal.

Artists and video game designers often construct 2D animations using libraries of sprites---textured patches of objects and characters. We propose a deep learning approach that decomposes sprite-based video animations into a disentangled representation of recurring graphic elements in a self-supervised manner. By jointly learning a dictionary of possibly transparent patches and training a network that places them onto a canvas, we deconstruct sprite-based content into a sparse, consistent, and explicit representation that can be easily used in downstream tasks, like editing or analysis. Our framework offers a promising approach for discovering recurring visual patterns in image collections without supervision.

High-quality denoising of Monte Carlo low-sample renderings remains a critical challenge for practical interactive ray tracing. We present a new learning-based denoiser that achieves state-of-the-art quality and runs at interactive rates. Our model processes individual path-traced samples with a lightweight neural network to extract per-pixel feature vectors. The rest of our pipeline operates in pixel space. We define a novel pairwise affinity over the features in a pixel neighborhood, from which we assemble dilated spatial kernels to filter the noisy radiance. Our denoiser is temporally stable thanks to two mechanisms. First, we keep a running average of the noisy radiance and intermediate features, using a per-pixel recursive filter with learned weights. Second, we use a small temporal kernel based on the pairwise affinity between features of consecutive frames. Our experiments show our new affinities lead to higher quality outputs than techniques with comparable computational costs, and better high-frequency details than kernel-predicting approaches. Our model matches or outperfoms state-of-the-art offline denoisers in the low-sample count regime (2--8 samples per pixel), and runs at interactive frame rates at 1080p resolution.

We introduce a neural relighting algorithm for captured indoors scenes, that allows interactive free-viewpoint navigation. Our method allows illumination to be changed synthetically, while coherently rendering cast shadows and complex glossy materials. We start with multiple images of the scene and a three-dimensional mesh obtained by multi-view stereo (MVS) reconstruction. We assume that lighting is well explained as the sum of a view-independent diffuse component and a view-dependent glossy term concentrated around the mirror reflection direction. We design a convolutional network around input feature maps that facilitate learning of an implicit representation of scene materials and illumination, enabling both relighting and free-viewpoint navigation. We generate these input maps by exploiting the best elements of both image-based and physically based rendering. We sample the input views to estimate diffuse scene irradiance, and compute the new illumination caused by user-specified light sources using path tracing. To facilitate the network's understanding of materials and synthesize plausible glossy reflections, we reproject the views and compute mirror images . We train the network on a synthetic dataset where each scene is also reconstructed with MVS. We show results of our algorithm relighting real indoor scenes and performing free-viewpoint navigation with complex and realistic glossy reflections, which so far remained out of reach for view-synthesis techniques.

Bursts of images exhibit significant self-similarity across both time and space. This motivates a representation of the kernels as linear combinations of a small set of basis elements. To this end, we introduce a novel basis prediction network that, given an input burst, predicts a set of global basis kernels --- shared within the image --- and the corresponding mixing coefficients --- which are specific to individual pixels. Compared to state-of-the-art techniques that output a large tensor of per-pixel spatiotemporal kernels, our formulation substantially reduces the dimensionality of the network output. This allows us to effectively exploit comparatively larger denoising kernels, achieving both significant quality improvements (over 1dB PSNR) and faster run-times over state-of-the-art methods.

We introduce a differentiable rasterizer that bridges the vector graphics and raster image domains, enabling powerful raster-based loss functions, optimization procedures, and machine learning techniques to edit and generate vector content. We observe that vector graphics rasterization is differentiable after pixel prefiltering. Our differentiable rasterizer offers two prefiltering options: an analytical prefiltering technique and a multisampling anti-aliasing technique. The analytical variant is faster but can suffer from artifacts such as conflation. The multisampling variant is still efficient, and can render high-quality images while computing unbiased gradients for each pixel with respect to curve parameters. We demonstrate that our rasterizer enables new applications, including a vector graphics editor guided by image metrics, a painterly rendering algorithm that fits vector primitives to an image by minimizing a deep perceptual loss function, new vector graphics editing algorithms that exploit well-known image processing methods such as seam carving, and deep generative models that generate vector content from raster-only supervision under a VAE or GAN training objective.

Collections of images under a single, uncontrolled illumination have enabled the rapid advancement of core computer vision tasks like classification, detection, and segmentation. But even with modern learning techniques, many inverse problems involving lighting and material understanding remain too severely ill-posed to be solved with single-illumination datasets. The data simply does not contain the necessary supervisory signals. Multi-illumination datasets are notoriously hard to capture, so the data is typically collected at small scale, in controlled environments, either using multiple light sources, or robotic gantries. This leads to image collections that are not representative of the variety and complexity of real world scenes. We introduce a new multi-illumination dataset of more than 1000 real scenes, each captured in high dynamic range and high resolution, under 25 lighting conditions. We demonstrate the richness of this dataset by training state-of-the-art models for three challenging applications: single-image illumination estimation, image relighting, and mixed-illuminant white balance.

We present a new algorithm to automatically schedule Halide programs for high-performance image processing and deep learning. We significantly improve upon the performance of previous methods, which considered a limited subset of schedules. We define a parameterization of possible schedules much larger than prior methods and use a variant of beam search to search over it. The search optimizes runtime predicted by a cost model based on a combination of new derived features and machine learning. We train the cost model by generating and featurizing hundreds of thousands of random programs and schedules. We show that this approach operates effectively with or without autotuning. It produces schedules which are on average almost twice as fast as the existing Halide autoscheduler without autotuning, or more than twice as fast with, and is the first automatic scheduling algorithm to significantly outperform human experts on average.

Denoising has proven to be useful to efficiently generate high-quality Monte Carlo renderings. Traditional pixel-based denoisers exploit summary statistics of a pixel's sample distributions, which discards much of the samples' information and limits their denoising power. On the other hand, sample-based techniques tend to be slow and have difficulties handling general transport scenarios. We present the first convolutional network that can learn to denoise Monte Carlo renderings directly from the samples. Learning the mapping between samples and images creates new challenges for the network architecture design: the order of the samples is arbitrary, and they should be treated in a permutation invariant manner. To address these challenges, we develop a novel kernel-predicting architecture that splats individual samples onto nearby pixels. Splatting is a natural solution to situations such as motion blur, depth-of-field and many light transport paths, where it is easier to predict which pixels a sample contributes to, rather than a gather approach that needs to figure out, for each pixel, which samples (or nearby pixels) are relevant. Compared to previous state-of-the-art methods, ours is robust to the severe noise of low-sample count images (e.g. 8 samples per pixel) and yields higher-quality results both visually and numerically. Our approach retains the generality and efficiency of pixel-space methods while enjoying the expressiveness and accuracy of the more complex sample-based approaches.

We propose the first learning-based algorithm that can relight images in a plausible and controllable manner given multiple views of an outdoor scene. In particular, we introduce a geometry-aware neural network that utilizes multiple geometry cues (normal maps, specular direction, etc.) and source and target shadow masks computed from a noisy proxy geometry obtained by multi-view stereo. Our model is a three-stage pipeline: two subnetworks refine the source and target shadow masks, and a third performs the final relighting. Furthermore, we introduce a novel representation for the shadow masks, which we call RGB shadow images. They reproject the colors from all views into the shadowed pixels and enable our network to cope with inacuraccies in the proxy and the non-locality of the shadow casting interactions. Acquiring large-scale multi-view relighting datasets for real scenes is challenging, so we train our network on photorealistic synthetic data. At train time, we also compute a noisy stereo-based geometric proxy, this time from the synthetic renderings. This allows us to bridge the gap between the real and synthetic domains. Our model generalizes well to real scenes. It can alter the illumination of drone footage, image-based renderings, textured mesh reconstructions, and even internet photo collections.

Gradient-based optimization has enabled dramatic advances in computational imaging through techniques like deep learning and nonlinear optimization. These methods require gradients not just of simple mathematical functions, but of general programs which encode complex transformations of images and graphical data. Unfortunately, practitioners have traditionally been limited to either hand-deriving gradients of complex computations, or composing programs from a limited set of coarse-grained operators in deep learning frameworks. At the same time, writing programs with the level of performance needed for imaging and deep learning is prohibitively difficult for most programmers. We extend the image processing language Halide with general reverse-mode automatic differentiation (AD), and the ability to automatically optimize the implementation of gradient computations. This enables automatic computation of the gradients of arbitrary Halide programs, at high performance, with little programmer effort. A key challenge is to structure the gradient code to retain parallelism. We define a simple algorithm to automatically schedule these pipelines, and show how Halide's existing scheduling primitives can express and extend the key AD optimization of "checkpointing." Using this new tool, we show how to easily define new neural network layers which automatically compile to high-performance GPU implementations, and how to solve nonlinear inverse problems from computational imaging. Finally, we show how differentiable programming enables dramatically improving the quality of even traditional, feed-forward image processing algorithms, blurring the distinction between classical and deep methods.

Performance is a critical challenge in mobile image processing. Given a reference imaging pipeline, or even human-adjusted pairs of images, we seek to reproduce the enhancements and enable real-time evaluation. For this, we introduce a new neural network architecture inspired by bilateral grid processing and local affine color transforms. Using pairs of input/output images, we train a convolutional neural network to predict the coefficients of a locally-affine model in bilateral space. Our architecture learns to make local, global, and content-dependent decisions to approximate the desired image transformation. At runtime, the neural network consumes a low-resolution version of the input image, produces a set of affine transformations in bilateral space, upsamples those transformations in an edge-preserving fashion using a new slicing node, and then applies those upsampled transformations to the full-resolution image. Our algorithm processes high-resolution images on a smartphone in milliseconds, provides a real-time viewfinder at 1080p resolution, and matches the quality of state-of-the-art approximation techniques on a large class of image operators. Unlike previous work, our model is trained off-line from data and therefore does not require access to the original operator at runtime. This allows our model to learn complex, scene-dependent transformations for which no reference implementation is available, such as the photographic edits of a human retoucher.

Demosaicking and denoising are the key first stages of the digital imaging pipeline but they are also a severely ill-posed problem that infers three color values per pixel from a single noisy measurement. Earlier methods rely on hand-crafted filters or priors and still exhibit disturbing visual artifacts in hard cases such as moiré or thin edges. We introduce a new data-driven approach for these challenges: we train a deep neural network on a large corpus of images instead of using hand-tuned filters. While deep learning has shown great success, its naive application using existing training datasets does not give satisfactory results for our problem because these datasets lack hard cases. To create a better training set, we present metrics to identify difficult patches and techniques for mining community photographs for such patches. Our experiments show that this network and training procedure outperform state-of-the-art both on noisy and noise-free data. Furthermore, our algorithm is an order of magnitude faster than the previous best performing techniques.

Cloud image processing is often proposed as a solution to the limited computing power and battery life of mobile devices: it allows complex algorithms to run on powerful servers with virtually unlimited energy supply. Unfortunately, this overlooks the time and energy cost of uploading the input and downloading the output images. When transfer overhead is accounted for, processing images on a remote server becomes less attractive and many applications do not benefit from cloud offloading. We aim to change this in the case of image enhancements that preserve the overall content of an image. Our key insight is that, in this case, the server can compute and transmit a description of the transformation from input to output, which we call a transform recipe. At equivalent quality, our recipes are much more compact than JPEG images: this reduces the client's download. Furthermore, recipes can be computed from highly compressed inputs which significantly reduces the data uploaded to the server. The client reconstructs a high-fidelity approximation of the output by applying the recipe to its local high-quality input. We demonstrate our results on 168 images and 10 image processing applications, showing that our recipes form a compact representation for a diverse set of image filters. With an equivalent transmission budget, they provide higher-quality results than JPEG-compressed input/output images, with a gain of the order of 10 dB in many cases. We demonstrate the utility of recipes on a mobile phone by profiling the energy consumption and latency for both local and cloud computation: a transform recipe-based pipeline runs 2--4x faster and uses 2--7x less energy than local or naive cloud computation.