Neural rendering provides a fundamentally new way to render photorealistic images. Similar to traditional light-baking methods, neural rendering utilizes neural networks to bake representations of scenes, materials, and lights into latent vectors learned from path-tracing ground truths. However, existing neural rendering algorithms typically use G-buffers to provide position, normal, and texture information about scenes. These are prone to occlusion by transparent surfaces, leading to distortion and loss of detail in rendered images. To address this limitation, we propose a novel neural rendering pipeline that accurately renders the scene behind transparent surfaces with global illumination and variable scenes. Our method separates the G-buffers for opaque and transparent objects, retaining G-buffer information behind transparent objects. Additionally, to render transparent objects with permutation invariance, we have designed a new permutation-invariant neural blending function. We have integrated our algorithm into an efficient custom renderer, achieving real-time performance. Our results show that our method is capable of rendering photorealistic images for variable scenes and viewpoints, accurately capturing complex transparent structures along with global illumination. Our renderer can achieve real-time performance (256 × 256 at 63 frames/s and 512 × 512 at 32 frames/s) for scenes with multiple variable transparent objects.

Game Description Language (GDL) provides a standardized way to express diverse games in a machine-readable format, enabling automated game simulation, and evaluation. While previous research has explored game description generation using search-based methods, generating GDL descriptions from natural language remains a challenging task. This paper presents a novel framework that leverages Large Language Models (LLMs) to generate grammatically accurate game descriptions from natural language. Our approach consists of two stages: first, we gradually generate a minimal grammar based on GDL specifications; second, we iteratively improve the game description through grammar-guided generation. Our framework employs a specialized parser that identifies valid subsequences and candidate symbols from LLM responses, enabling gradual refinement of the output to ensure grammatical correctness. Experimental results demonstrate that our iterative improvement approach significantly outperforms baseline methods that directly use LLM outputs. Our code is available athttps://github.com/tsunehiko/ggdg

Creating high-quality line art in a fast and controlled manner plays a crucial role in anime production and concept design. We present DoodleAssist, an interactive and progressive line art generation system controlled by sketches and prompts, which helps both experts and novices concretize their design intentions or explore possibilities. Built upon a controllable diffusion model, our system performs progressive generation based on the last generated line art, synthesizing regions corresponding to drawn or modified strokes while keeping the remaining ones unchanged. To facilitate this process, we propose a latent distribution alignment mechanism to enhance the transition between the two regions and allow seamless blending, thereby alleviating issues of region incoherence and line discontinuity. Finally, we also build a user interface that allows the convenient creation of line art through interactive sketching and prompts. Qualitative and quantitative comparisons against existing approaches and an in-depth user study demonstrate the effectiveness and usability of our system. Our system can benefit various applications such as anime concept design, drawing assistant, and creativity support for children.

Game Description Generation (GDG) is the task of generating a game description written in a Game Description Language (GDL) from natural language text. Previous studies have explored generation methods leveraging the contextual understanding capabilities of Large Language Models (LLMs); however, accurately reproducing the game features of the game descriptions remains a challenge. In this paper, we propose reinforcement learning-based fine-tuning of LLMs for GDG (RLGDG). Our training method simultaneously improves grammatical correctness and fidelity to game concepts by introducing both grammar rewards and concept rewards. Furthermore, we adopt a two-stage training strategy where Reinforcement Learning (RL) is applied following Supervised Fine-Tuning (SFT). Experimental results demonstrate that our proposed method significantly outperforms baseline methods using SFT alone.

We introduce MarkupDM, a multimodal markup document model that represents graphic design as an interleaved multimodal document consisting of both markup language and images. Unlike existing holistic approaches that rely on an element-by-attribute grid representation, our representation accommodates variable-length elements, type-dependent attributes, and text content. Inspired by fill-in-the-middle training in code generation, we train the model to complete the missing part of a design document from its surrounding context, allowing it to treat various design tasks in a unified manner. Our model also supports image generation by predicting discrete image tokens through a specialized tokenizer with support for image transparency. We evaluate MarkupDM on three tasks, attribute value, image, and text completion, and demonstrate that it can produce plausible designs consistent with the given context. To further illustrate the flexibility of our approach, we evaluate our approach on a new instruction-guided design completion task where our instruction-tuned MarkupDM compares favorably to state-of-the-art image editing models, especially in textual completion. These findings suggest that multimodal language models with our document representation can serve as a versatile foundation for broad design automation.

There exists a large number of old films that have not only artistic value but also historical significance. However, due to the degradation of analogue medium over time, old films often suffer from various deteriorations that make it difficult to restore them with existing approaches. In this work, we proposed a novel framework called Recursive Recurrent Transformer Network (RRTN) which is specifically designed for restoring degraded old films. Our approach introduces several key advancements, including a more accurate film noise mask estimation method, the utilization of second-order grid propagation and flow-guided deformable alignment, and the incorporation of a recursive structure to further improve the removal of challenging film noise. Through qualitative and quantitative evaluations, our approach demonstrates superior performance compared to existing approaches, effectively improving the restoration for difficult film noises that cannot be perfectly handled by existing approaches. The code and model are available at https://github.com/mountln/RRTN-old-film-restoration.

Abstract Neural rendering bakes global illumination and other computationally costly effects into the weights of a neural network, allowing to efficiently synthesize photorealistic images without relying on path tracing. In neural rendering approaches, G‐buffers obtained from rasterization through direct rendering provide information regarding the scene such as position, normal, and textures to the neural network, achieving accurate and stable rendering quality in real‐time. However, due to the use of G‐buffers, existing methods struggle to accurately render transparency and refraction effects, as G‐buffers do not capture any ray information from multiple light ray bounces. This limitation results in blurriness, distortions, and loss of detail in rendered images that contain transparency and refraction, and is particularly notable in scenes with refracted objects that have high‐frequency textures. In this work, we propose a neural network architecture to encode critical rendering information, including texture coordinates from refracted rays, and enable reconstruction of high‐frequency textures in areas with refraction. Our approach is able to achieve accurate refraction rendering in challenging scenes with a diversity of overlapping transparent objects. Experimental results demonstrate that our method can interactively render high quality refraction effects with global illumination, unlike existing neural rendering approaches. Our code can be found at https://github.com/ziyangz5/CrystalNet

We present a novel framework for the multi-domain synthesis of artworks from semantic layouts. One of the main limitations of this challenging task is the lack of publicly available segmentation datasets for art synthesis. To address this problem, we propose a dataset called ArtSem that contains 40,000 images of artwork from four different domains, with their corresponding semantic label maps. We first extracted semantic maps from landscape photography and used a conditional generative adversarial network (GAN)-based approach for generating high-quality artwork from semantic maps without requiring paired training data. Furthermore, we propose an artwork-synthesis model using domain-dependent variational encoders for high-quality multi-domain synthesis. Subsequently, the model was improved and complemented with a simple but effective normalization method based on jointly normalizing semantics and style, which we call spatially style-adaptive normalization (SSTAN). Compared to the previous methods, which only take semantic layout as the input, our model jointly learns style and semantic information representation, improving the generation quality of artistic images. These results indicate that our model learned to separate the domains in the latent space. Thus, we can perform fine-grained control of the synthesized artwork by identifying hyperplanes that separate the different domains. Moreover, by combining the proposed dataset and approach, we generated user-controllable artworks of higher quality than that of existing approaches, as corroborated by quantitative metrics and a user study.

We introduce an algorithm for sketch vectorization with state-of-the-art accuracy and capable of handling complex sketches. We approach sketch vectorization as a surface extraction task from an unsigned distance field, which is implemented using a two-stage neural network and a dual contouring domain post processing algorithm. The first stage consists of extracting unsigned distance fields from an input raster image. The second stage consists of an improved neural dual contouring network more robust to noisy input and more sensitive to line geometry. To address the issue of under-sampling inherent in grid-based surface extraction approaches, we explicitly predict undersampling and keypoint maps. These are used in our post-processing algorithm to resolve sharp features and multi-way junctions. The keypoint and undersampling maps are naturally controllable, which we demonstrate in an interactive topology refinement interface. Our proposed approach produces far more accurate vectorizations on complex input than previous approaches with efficient running time.

Controllable layout generation aims at synthesizing plausible arrangement of element bounding boxes with optional constraints, such as type or position of a specific element. In this work, we try to solve a broad range of layout generation tasks in a single model that is based on discrete state-space diffusion models. Our model, named Lay-outDM, naturally handles the structured layout data in the discrete representation and learns to progressively infer a noiseless layout from the initial input, where we model the layout corruption process by modality-wise discrete diffusion. For conditional generation, we propose to inject layout constraints in the form of masking or logit adjustment during inference. We show in the experiments that our Lay-outDM successfully generates high-quality layouts and outperforms both task-specific and task-agnostic baselines on several layout tasks.11Please find the code and models at: https://cyberagentailab.github.io/layout-drn.

Creative workflows for generating graphical documents involve complex inter-related tasks, such as aligning elements, choosing appropriate fonts, or employing aesthetically harmonious colors. In this work, we attempt at building a holistic model that can jointly solve many different design tasks. Our model, which we denote by FlexDM, treats vector graphic documents as a set of multi-modal elements, and learns to predict masked fields such as element type, position, styling attributes, image, or text, using a unified architecture. Through the use of explicit multitask learning and in-domain pre-training, our model can better capture the multi-modal relationships among the different document fields. Experimental results corroborate that our single FlexDM is able to successfully solve a multitude of different design tasks, while achieving performance that is competitive with task-specific and costly baselines.11Please find the code and models at: https://cyberagentailab.github.io/flex-dm

We present a novel framework for rectifying occlusions and distortions in degraded texture samples from natural images. Traditional texture synthesis approaches focus on generating textures from pristine samples, which necessitate meticulous preparation by humans and are often unattainable in most natural images. These challenges stem from the frequent occlusions and distortions of texture samples in natural images due to obstructions and variations in object surface geometry. To address these issues, we propose a framework that synthesizes holistic textures from degraded samples in natural images, extending the applicability of exemplar-based texture synthesis techniques. Our framework utilizes a conditional Latent Diffusion Model (LDM) with a novel occlusion-aware latent transformer. This latent transformer not only effectively encodes texture features from partially-observed samples necessary for the generation process of the LDM, but also explicitly captures long-range dependencies in samples with large occlusions. To train our model, we introduce a method for generating synthetic data by applying geometric transformations and free-form mask generation to clean textures. Experimental results demonstrate that our framework significantly outperforms existing methods both quantitatively and quantitatively. Furthermore, we conduct comprehensive ablation studies to validate the different components of our proposed framework. Results are corroborated by a perceptual user study which highlights the efficiency of our proposed approach.

Color is a critical design factor for web pages, affecting important factors such as viewer emotions and the overall trust and satisfaction of a website. Effective coloring requires design knowledge and expertise, but if this process could be automated through data-driven modeling, efficient exploration and alternative workflows would be possible. However, this direction remains underexplored due to the lack of a formalization of the web page colorization problem, datasets, and evaluation protocols. In this work, we propose a new dataset consisting of e-commerce mobile web pages in a tractable format, which are created by simplifying the pages and extracting canonical color styles with a common web browser. The web page colorization problem is then formalized as a task of estimating plausible color styles for a given web page content with a given hierarchical structure of the elements. We present several Transformer-based methods that are adapted to this task by prepending structural message passing to capture hierarchical relation-ships between elements. Experimental results, including a quantitative evaluation designed for this task, demonstrate the advantages of our methods over statistical and image colorization methods. The code is available at https://github.com/CyberAgentAILab/webcolor.

Abstract Although digital painting has advanced much in recent years, there is still a significant divide between physically drawn paintings and purely digitally drawn paintings. These differences arise due to the physical interactions between the brush, ink, and paper, which are hard to emulate in the digital domain. Most ink painting approaches have focused on either using heuristics or physical simulation to attempt to bridge the gap between digital and analog, however, these approaches are still unable to capture the diversity of painting effects, such as ink fading or blotting, found in the real world. In this work, we propose a data‐driven approach to generate ink paintings based on a semi‐automatically collected high‐quality real‐world ink painting dataset. We use a multi‐camera robot‐based setup to automatically create a diversity of ink paintings, which allows for capturing the entire process in high resolution, including capturing detailed brush motions and drawing results. To ensure high‐quality capture of the painting process, we calibrate the setup and perform occlusion‐aware blending to capture all the strokes in high resolution in a robust and efficient way. Using our new dataset, we propose a recursive deep learning‐based model to reproduce the ink paintings stroke by stroke while capturing complex ink painting effects such as bleeding and mixing. Our results corroborate the fidelity of the proposed approach to real hand‐drawn ink paintings in comparison with existing approaches. We hope the availability of our dataset will encourage new research on digital realistic ink painting techniques.

Dealing with unstructured complex patterns provides a challenge to existing reinforcement patterns. In this research, we propose a new model to overcome the difficulty in challenging danmaku games. Touhou Project is one of the bestknown games in the bullet hell genre also known as danmaku, where a player has to dodge complex patterns of bullets on the screen. Furthermore, the agent needs to react to the environment in real-time, which made existing methods having difficulties processing the high-volume data of objects; bullets, enemies, etc. We introduce an environment for the Touhou Project game‘東方花映塚~Phantasmagoria of Flower View.’ which manipulates the memory of the running game and enables to control the character. However, the game state information consists of unstructured and unordered data not amenable for training existing reinforcement learning models, as they are not invariant to order changes in the input. To overcome this issue, we propose a new pooling-based reinforcement learning approach that is able to handle permutation invariant inputs by extracting abstract values and merging them in an order-independent way. Experimental results corroborate the effectiveness of our approach which shows significantly increased scores compared to existing baseline approaches.

Roguelike games are a challenging environment for Reinforcement Learning (RL) algorithms due to having to restart the game from the beginning when losing, randomized procedural generation, and proper use of in-game items being essential to success. While recent research has proposed roguelike environments for RL algorithms and proposed models to handle this challenging task, to the best of our knowledge, none have dealt with the elephant in the room, i.e., handling of items. Items play a fundamental role in roguelikes and are acquired during gameplay. However, being an unordered set with a non-fixed amount of elements which form part of the action space, it is not straightforward to incorporate them into an RL framework. In this work, we tackle the issue of having unordered sets be part of the action space and propose an attention-based mechanism that can select and deal with item-based actions. We also propose a model that can handle complex actions and items through a meta action framework and evaluate them on the challenging game of NetHack. Experimental results show that our approach is able to significantly outperform existing approaches.

Flat filling is a critical step in digital artistic content creation with the objective of filling line arts with flat colors. We present a deep learning framework for user-guided line art flat filling that can compute the "influence areas" of the user color scribbles, i.e., the areas where the user scribbles should propagate and influence. This framework explicitly controls such scribble influence areas for artists to manipulate the colors of image details and avoid color leakage/contamination between scribbles, and simultaneously, leverages data-driven color generation to facilitate content creation. This framework is based on a Split Filling Mechanism (SFM), which first splits the user scribbles into individual groups and then independently processes the colors and influence areas of each group with a Convolutional Neural Network (CNN). Learned from more than a million illustrations, the framework can estimate the scribble influence areas in a content-aware manner, and can smartly generate visually pleasing colors to assist the daily works of artists. We show that our proposed framework is easy to use, allowing even amateurs to obtain professional-quality results on a wide variety of line arts.

It is common in graphic design humans visually arrange various elements according to their design intent and semantics. For example, a title text almost always appears on top of other elements in a document. In this work, we generate graphic layouts that can flexibly incorporate such design semantics, either specified implicitly or explicitly by a user. We optimize using the latent space of an off-the-shelf layout generation model, allowing our approach to be complementary to and used with existing layout generation models. Our approach builds on a generative layout model based on a Transformer architecture, and formulates the layout generation as a constrained optimization problem where design constraints are used for element alignment, overlap avoidance, or any other user-specified relationship. We show in the experiments that our approach is capable of generating realistic layouts in both constrained and unconstrained generation tasks with a single model. The code is available at https://github.com/ktrk115/const_layout.

Abstract Web pages have become fundamental in conveying information for companies and individuals, yet designing web page layouts remains a challenging task for inexperienced individuals despite web builders and templates. Visual containment, in which elements are grouped together and placed inside container elements, is an efficient design strategy for organizing elements in a limited display, and is widely implemented in most web page designs. Yet, visual containment has not been explicitly addressed in the research on generating layouts from scratch, which may be due to the lack of hierarchical structure. In this work, we represent such visual containment as a layout tree, and formulate the layout design task as a hierarchical optimization problem. We first estimate the layout tree from a given a set of elements, which is then used to compute tree‐aware energies corresponding to various desirable design properties such as alignment or spacing. Using an optimization approach also allows our method to naturally incorporate user intentions and create an interactive web design application. We obtain a dataset of diverse and popular real‐world web designs to optimize and evaluate various aspects of our method. Experimental results show that our method generates better quality layouts compared to the baseline method.

Vector line art plays an important role in graphic design, however, it is tedious to manually create. We introduce a general framework to produce line drawings from a wide variety of images, by learning a mapping from raster image space to vector image space. Our approach is based on a recurrent neural network that draws the lines one by one. A differentiable rasterization module allows for training with only supervised raster data. We use a dynamic window around a virtual pen while drawing lines, implemented with a proposed aligned cropping and differentiable pasting modules. Furthermore, we develop a stroke regularization loss that encourages the model to use fewer and longer strokes to simplify the resulting vector image. Ablation studies and comparisons with existing methods corroborate the efficiency of our approach which is able to generate visually better results in less computation time, while generalizing better to a diversity of images and applications.

We present a lightweight post-processing method to refine the semantic segmentation results of point cloud sequences. Most existing methods usually segment frame by frame and encounter the inherent ambiguity of the problem: based on a measurement in a single frame, labels are sometimes difficult to predict even for humans. To remedy this problem, we propose to explicitly train a network to refine these results predicted by an existing segmentation method. The network, which we call the P2Net, learns the consistency constraints between "coincident" points from consecutive frames after registration. We evaluate the proposed post-processing method both qualitatively and quantitatively on the SemanticKITTI dataset that consists of real outdoor scenes. The effectiveness of the proposed method is validated by comparing the results predicted by two representative networks with and without the refinement by the post-processing network. Specifically, qualitative visualization validates the key idea that labels of the points that are difficult to predict can be corrected with P2Net. Quantitatively, overall mIoU is improved from 10.5% to 11.7% for PointNet [1] and from 10.8% to 15.9% for PointNet++ [2].

We present an algorithm to generate digital painting lighting effects from a single image. Our algorithm is based on a key observation: Artists use many overlapping strokes to paint lighting effects, i.e., pixels with dense stroke history tend to gather more illumination strokes. Based on this observation, we design an algorithm to both estimate the density of strokes in a digital painting using color geometry and then generate novel lighting effects by mimicking artists’ coarse-to-fine workflow. Coarse lighting effects are first generated using a wave transform and then retouched according to the stroke density of the original illustrations into usable lighting effects. Our algorithm is content-aware, with generated lighting effects naturally adapting to image structures, and can be used as an interactive tool to simplify current labor-intensive workflows for generating lighting effects for digital and matte paintings. In addition, our algorithm can also produce usable lighting effects for photographs or three-dimensional rendered images. We evaluate our approach with both an in-depth qualitative and a quantitative analysis that includes a perceptual user study. Results show that our proposed approach is not only able to produce favorable lighting effects with respect to existing approaches, but also that it is able to significantly reduce the needed interaction time.

The remastering of vintage film comprises of a diversity of sub-tasks including super-resolution, noise removal, and contrast enhancement which aim to restore the deteriorated film medium to its original state. Additionally, due to the technical limitations of the time, most vintage film is either recorded in black and white, or has low quality colors, for which colorization becomes necessary. In this work, we propose a single framework to tackle the entire remastering task semi-interactively. Our work is based on temporal convolutional neural networks with attention mechanisms trained on videos with data-driven deterioration simulation. Our proposed source-reference attention allows the model to handle an arbitrary number of reference color images to colorize long videos without the need for segmentation while maintaining temporal consistency. Quantitative analysis shows that our framework outperforms existing approaches, and that, in contrast to existing approaches, the performance of our framework increases with longer videos and more reference color images.

We present an interactive approach for inking , which is the process of turning a pencil rough sketch into a clean line drawing. The approach, which we call the Smart Inker , consists of several "smart" tools that intuitively react to user input, while guided by the input rough sketch, to efficiently and naturally connect lines, erase shading, and fine-tune the line drawing output. Our approach is data-driven: the tools are based on fully convolutional networks, which we train to exploit both the user edits and inaccurate rough sketch to produce accurate line drawings, allowing high-performance interactive editing in real-time on a variety of challenging rough sketch images. For the training of the tools, we developed two key techniques: one is the creation of training data by simulation of vague and quick user edits; the other is a line normalization based on learning from vector data. These techniques, in combination with our sketch-specific data augmentation, allow us to train the tools on heterogeneous data without actual user interaction. We validate our approach with an in-depth user study, comparing it with professional illustration software, and show that our approach is able to reduce inking time by a factor of 1.8X, while improving the results of amateur users.

We present an integral framework for training sketch simplification networks that convert challenging rough sketches into clean line drawings. Our approach augments a simplification network with a discriminator network, training both networks jointly so that the discriminator network discerns whether a line drawing is real training data or the output of the simplification network, which, in turn, tries to fool it. This approach has two major advantages: first, because the discriminator network learns the structure in line drawings, it encourages the output sketches of the simplification network to be more similar in appearance to the training sketches. Second, we can also train the networks with additional unsupervised data: by adding rough sketches and line drawings that are not corresponding to each other, we can improve the quality of the sketch simplification. Thanks to a difference in the architecture, our approach has advantages over similar adversarial training approaches in stability of training and the aforementioned ability to utilize unsupervised training data. We show how our framework can be used to train models that significantly outperform the state of the art in the sketch simplification task, despite using the same architecture for inference. We also present an approach to optimize for a single image, which improves accuracy at the cost of additional computation time. Finally, we show that, using the same framework, it is possible to train the network to perform the inverse problem, i.e., convert simple line sketches into pencil drawings, which is not possible using the standard mean squared error loss. We validate our framework with two user tests, in which our approach is preferred to the state of the art in sketch simplification 88.9% of the time.

We propose a novel data-driven approach for automatically detecting and completing gaps in line drawings with a Convolutional Neural Network. In the case of existing inpainting approaches for natural images, masks indicating the missing regions are generally required as input. Here, we show that line drawings have enough structures that can be learned by the CNN to allow automatic detection and completion of the gaps without any such input. Thus, our method can find the gaps in line drawings and complete them without user interaction. Furthermore, the completion realistically conserves thickness and curvature of the line segments. All the necessary heuristics for such realistic line completion are learned naturally from a dataset of line drawings, where various patterns of line completion are generated on the fly as training pairs to improve the model generalization. We evaluate our method qualitatively on a diverse set of challenging line drawings and also provide quantitative results with a user study, where it significantly outperforms the state of the art.

Finding a product in the fashion world can be a daunting task. Everyday, e-commerce sites are updating with thousands of images and their associated metadata (textual information), deepening the problem, akin to finding a needle in a haystack. In this paper, we leverage both the images and textual metadata and propose a joint multi-modal embedding that maps both the text and images into a common latent space. Distances in the latent space correspond to similarity between products, allowing us to effectively perform retrieval in this latent space, which is both efficient and accurate. We train this embedding using large-scale real world e-commerce data by both minimizing the similarity between related products and using auxiliary classification networks to that encourage the embedding to have semantic meaning. We compare against existing approaches and show significant improvements in retrieval tasks on a large-scale e-commerce dataset. We also provide an analysis of the different metadata.

We present a novel approach for image completion that results in images that are both locally and globally consistent. With a fully-convolutional neural network, we can complete images of arbitrary resolutions by filling-in missing regions of any shape. To train this image completion network to be consistent, we use global and local context discriminators that are trained to distinguish real images from completed ones. The global discriminator looks at the entire image to assess if it is coherent as a whole, while the local discriminator looks only at a small area centered at the completed region to ensure the local consistency of the generated patches. The image completion network is then trained to fool the both context discriminator networks, which requires it to generate images that are indistinguishable from real ones with regard to overall consistency as well as in details. We show that our approach can be used to complete a wide variety of scenes. Furthermore, in contrast with the patch-based approaches such as PatchMatch, our approach can generate fragments that do not appear elsewhere in the image, which allows us to naturally complete the images of objects with familiar and highly specific structures, such as faces.

We propose a novel approach for learning features from weakly-supervised data by joint ranking and classification. In order to exploit data with weak labels, we jointly train a feature extraction network with a ranking loss and a classification network with a cross-entropy loss. We obtain high-quality compact discriminative features with few parameters, learned on relatively small datasets without additional annotations. This enables us to tackle tasks with specialized images not very similar to the more generic ones in existing fully-supervised datasets. We show that the resulting features in combination with a linear classifier surpass the state-of-the-art on the Hipster Wars dataset despite using features only 0.3% of the size. Our proposed features significantly outperform those obtained from networks trained on ImageNet, despite being 32 times smaller (128 single-precision floats), trained on noisy and weakly-labeled data, and using only 1.5% of the number of parameters.

We propose a method for understanding a room from a single spherical image, i.e., reconstructing and identifying structural planes forming the ceiling, the floor, and the walls in a room. A spherical image records the light that falls onto a single viewpoint from all directions and does not require correlating geometrical information from multiple images, which facilitates robust and precise reconstruction of the room structure. In our method, we detect line segments from a given image, and classify them into two groups: segments that form the boundaries of the structural planes and those that do not. We formulate this problem as a higher-order energy minimization problem that combines the various measures of likelihood that one, two, or three line segments are part of the boundary. We minimize the energy with graph cuts to identify segments forming boundaries, from which we estimate structural the planes in 3D. Experimental results on synthetic and real images confirm the effectiveness of the proposed method.

We present an approach for the detection of buildings in multispectral satellite images. Unlike 3-channel RGB images, satellite imagery contains additional channels corresponding to different wavelengths. Approaches that do not use all channels are unable to fully exploit these images for optimal performance. Furthermore, care must be taken due to the large bias in classes, e.g., most of the Earth is covered in water and thus it will be dominant in the images. Our approach consists of training a Convolutional Neural Network (CNN) from scratch to classify multispectral image patches taken by satellites as whether or not they belong to a class of buildings. We then adapt the classification network to detection by converting the fully-connected layers of the network to convolutional layers, which allows the network to process images of any resolution. The dataset bias is compensated by subsampling negatives and tuning the detection threshold for optimal performance. We have constructed a new dataset using images from the Landsat 8 satellite for detecting solar power plants and show our approach is able to significantly outperform the state-of-the-art. Furthermore, we provide an indepth evaluation of the seven different spectral bands provided by the satellite images and show it is critical to combine them to obtain good results.

We propose a new superpixel algorithm based on exploiting the boundary information of an image, as objects in images can generally be described by their boundaries. Our proposed approach initially estimates the boundaries and uses them to place superpixel seeds in the areas in which they are more dense. Afterwards, we minimize an energy function in order to expand the seeds into full superpixels. In addition to standard terms such as color consistency and compactness, we propose using the geodesic distance which concentrates small superpixels in regions of the image with more information, while letting larger superpixels cover more homogeneous regions. By both improving the initialization using the boundaries and coherency of the superpixels with geodesic distances, we are able to maintain the coherency of the image structure with fewer superpixels than other approaches. We show the resulting algorithm to yield smaller Variation of Information metrics in seven different datasets while maintaining Undersegmentation Error values similar to the state-of-the-art methods.

Ariadna Quattoni, Arnau Ramisa, Pranava Swaroop Madhyastha, Edgar Simo-Serra, Francesc Moreno-Noguer. Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2016.

We present a novel technique to automatically colorize grayscale images that combines both global priors and local image features. Based on Convolutional Neural Networks, our deep network features a fusion layer that allows us to elegantly merge local information dependent on small image patches with global priors computed using the entire image. The entire framework, including the global and local priors as well as the colorization model, is trained in an end-to-end fashion. Furthermore, our architecture can process images of any resolution, unlike most existing approaches based on CNN. We leverage an existing large-scale scene classification database to train our model, exploiting the class labels of the dataset to more efficiently and discriminatively learn the global priors. We validate our approach with a user study and compare against the state of the art, where we show significant improvements. Furthermore, we demonstrate our method extensively on many different types of images, including black-and-white photography from over a hundred years ago, and show realistic colorizations.

In this paper, we present a novel technique to simplify sketch drawings based on learning a series of convolution operators. In contrast to existing approaches that require vector images as input, we allow the more general and challenging input of rough raster sketches such as those obtained from scanning pencil sketches. We convert the rough sketch into a simplified version which is then amendable for vectorization. This is all done in a fully automatic way without user intervention. Our model consists of a fully convolutional neural network which, unlike most existing convolutional neural networks, is able to process images of any dimensions and aspect ratio as input, and outputs a simplified sketch which has the same dimensions as the input image. In order to teach our model to simplify, we present a new dataset of pairs of rough and simplified sketch drawings. By leveraging convolution operators in combination with efficient use of our proposed dataset, we are able to train our sketch simplification model. Our approach naturally overcomes the limitations of existing methods, e.g. , vector images as input and long computation time; and we show that meaningful simplifications can be obtained for many different test cases. Finally, we validate our results with a user study in which we greatly outperform similar approaches and establish the state of the art in sketch simplification of raster images.

In this paper, we analyze the fashion of clothing of a large social website. Our goal is to learn and predict how fashionable a person looks on a photograph and suggest subtle improvements the user could make to improve her/his appeal. We propose a Conditional Random Field model that jointly reasons about several fashionability factors such as the type of outfit and garments the user is wearing, the type of the user, the photograph's setting (e.g., the scenery behind the user), and the fashionability score. Importantly, our model is able to give rich feedback back to the user, conveying which garments or even scenery she/he should change in order to improve fashionability. We demonstrate that our joint approach significantly outperforms a variety of intelligent baselines. We additionally collected a novel heterogeneous dataset with 144,169 user posts containing diverse image, textual and meta information which can be exploited for our task. We also provide a detailed analysis of the data, showing different outfit trends and fashionability scores across the globe and across a span of 6 years.

Deep learning has revolutionalized image-level tasks such as classification, but patch-level tasks, such as correspondence, still rely on hand-crafted features, e.g. SIFT. In this paper we use Convolutional Neural Networks (CNNs) to learn discriminant patch representations and in particular train a Siamese network with pairs of (non-)corresponding patches. We deal with the large number of potential pairs with the combination of a stochastic sampling of the training set and an aggressive mining strategy biased towards patches that are hard to classify. By using the L2 distance during both training and testing we develop 128-D descriptors whose euclidean distances reflect patch similarity, and which can be used as a drop-in replacement for any task involving SIFT. We demonstrate consistent performance gains over the state of the art, and generalize well against scaling and rotation, perspective transformation, non-rigid deformation, and illumination changes. Our descriptors are efficient to compute and amenable to modern GPUs, and are publicly available.

We propose a robust and efficient method to estimate the pose of a camera with respect to complex 3D textured models of the environment that can potentially contain more than 100; 000 points. To tackle this problem we follow a top down approach where we combine high-level deep network classifiers with low level geometric approaches to come up with a solution that is fast, robust and accurate. Given an input image, we initially use a pre-trained deep network to compute a rough estimation of the camera pose. This initial estimate constrains the number of 3D model points that can be seen from the camera viewpoint. We then establish 3D-to-2D correspondences between these potentially visible points of the model and the 2D detected image features. Accurate pose estimation is finally obtained from the 2D-to-3D correspondences using a novel PnP algorithm that rejects outliers without the need to use a RANSAC strategy, and which is between 10 and 100 times faster than other methods that use it. Two real experiments dealing with very large and complex 3D models demonstrate the effectiveness of the approach.

We introduce a novel approach to automatically recover 3D human pose from a single image. Most previous work follows a pipelined approach: initially, a set of 2D features such as edges, joints or silhouettes are detected in the image, and then these observations are used to infer the 3D pose. Solving these two problems separately may lead to erroneous 3D poses when the feature detector has performed poorly. In this paper, we address this issue by jointly solving both the 2D detection and the 3D inference problems. For this purpose, we propose a Bayesian framework that integrates a generative model based on latent variables and discriminative 2D part detectors based on HOGs, and perform inference using evolutionary algorithms. Real experimentation demonstrates competitive results, and the ability of our methodology to provide accurate 2D and 3D pose estimations even when the 2D detectors are inaccurate.

Markerless 3D human pose detection from a single image is a severely underconstrained problem because different 3D poses can have similar image projections. In order to handle this ambiguity, current approaches rely on prior shape models that can only be correctly adjusted if 2D image features are accurately detected. Unfortunately, although current 2D part detector algorithms have shown promising results, they are not yet accurate enough to guarantee a complete disambiguation of the 3D inferred shape. In this paper, we introduce a novel approach for estimating 3D human pose even when observations are noisy. We propose a stochastic sampling strategy to propagate the noise from the image plane to the shape space. This provides a set of ambiguous 3D shapes, which are virtually undistinguishable from their image projections. Disambiguation is then achieved by imposing kinematic constraints that guarantee the resulting pose resembles a 3D human shape. We validate the method on a variety of situations in which state-of-the-art 2D detectors yield either inaccurate estimations or partly miss some of the body parts.