PUBLICATIONS

Point clouds are gaining prominence as a method for representing 3D shapes, but their irregular structure poses a challenge for deep learning methods. In this paper we propose CloudWalker, a novel method for learning 3D shapes using random walks. Previous works attempt to adapt Convolutional Neural Networks (CNNs) or impose a grid or mesh structure to 3D point clouds. This work presents a different approach for representing and learning the shape from a given point set. The key idea is to impose structure on the point set by multiple random walks through the cloud for exploring different regions of the 3D object. Then we learn a per-point and per-walk representation and aggregate multiple walk predictions at inference. Our approach achieves state-ofthe-art results for two 3D shape analysis tasks: classification and retrieval.

Archaeology is an intriguing domain for computer vision. It suffers not only from shortage in (labeled) data, but also from highly-hallenging data, which is often extremely abraded and damaged. This paper proposes a novel semi-supervised model for classification and retrieval of images of archaeological artifacts. This model utilizes unique data that exists in the domain — manual drawings made by special artists.These are used during training to implicitly transfer the domain knowledge from the drawings to their corresponding images, improving their classification results. We show that while learning how to classify, our model also learns how to generate drawings of the artifacts, an important documentation task, which is currently performed manually. Last but not least, we collected a new dataset of stamp-seals of the Southern Levant.

A polygonal mesh is the most-commonly used representation of surfaces in computer graphics. Therefore, it is not surprising that a number of mesh classification networks have recently been proposed. However, while adversarial attacks are wildly researched in 2D, the field of adversarial meshes is under explored. This paper proposes a novel, unified, and general adversarial attack, which leads to misclassification of several state-of-the-art mesh classification neural networks. Our attack approach is black-box, i.e. it has access only to the network’s predictions, but not to the network’s full architecture or gradients. The key idea is to train a network to imitate a given classification network. This is done by utilizing random walks along the mesh surface, which gather geometric information. These walks provide insight onto the regions of the mesh that are important for the correct prediction of the given classification network. These mesh regions are then modified more than other regions in order to attack the network in a manner that is barely visible to the naked eye.

This work focuses on object goal visual navigation, aiming at finding the location of an object from a given class, where in each step the agent is provided with an egocentric RGB image of the scene. We propose to learn the agent’s policy using a reinforcement learning algorithm. Our key contribution is a novel attention probability model for visual navigation tasks. This attention encodes semantic information about observed objects, as well as spatial information about their place. This combination of the “what” and the “where” allows the agent to navigate toward the sought-after object effectively. The attention model is shown to improve the agent’s policy and to achieve state-of-the-art results on commonly-used datasets.

Puzzle solving is a difficult problem in its own right, even when the pieces are all square and build up a natural image. But what if these ideal conditions do not hold? One such application domain is archaeology, where restoring an artifact from its fragments is highly important. From the point of view of computer vision, archaeological puzzle solving is very challenging, due to three additional difficulties: the fragments are of general shape; they are abraded, especially at the boundaries (where the strongest cues for matching should exist); and the domain of valid transformations between the pieces is continuous. The key contribution of this paper is a fully-automatic and general algorithm that addresses puzzle solving in this intriguing domain. We show that our state-of-the-art approach manages to correctly reassemble dozens of broken artifacts and frescoes.

Large-scale multi-label classification datasets are commonly, and perhaps inevitably, partially annotated. That is, only a small subset of labels are annotated per sample. Different methods for handling the missing labels induce different properties on the model and impact its accuracy. In this work, we analyze the partial labeling problem, then propose a solution based on two key ideas. First, un-annotated labels should be treated selectively according to two probability quantities: the class distribution in the overall dataset and the specific label likelihood for a given data sample. We propose to estimate the class distribution using a dedicated temporary model, and we show its improved efficiency over a naive estimation computed using the dataset’s partial annotations. Second, during the training of the target model, we emphasize the contribution of annotated labels over originally un-annotated labels by using a dedicated asymmetric loss. With our novel approach, we achieve state-of-the-art results on OpenImages dataset (e.g. reaching 87.3 mAP on V6). In addition, experiments conducted on LVIS and simulated-COCO demonstrate the effectiveness of our approach.

In recent years the amounts of personal photos captured increased significantly, giving rise to new challenges in multi-image understanding and high-level image understanding. Event recognition in personal photo albums presents one challenging scenario where life events are recognized from a disordered collection of images, including both relevant and irrelevant images. Event recognition in images also presents the challenge of high-level image understanding, as opposed to low-level image object classification. In absence of methods to analyze multiple inputs, previous methods adopted temporal mechanisms, including various forms of recurrent neural networks. However, their effective temporal window is local. In addition, they are not a natural choice given the disordered characteristic of photo albums. We address this gap with a tailor-made solution, combining the power of CNNs for image representation and transformers for album representation to perform global reasoning on image collection, offering a practical and efficient solution for photo albums event recognition. Our solution reaches state-of-the-art results on 3 prominent benchmarks, achieving above 90% mAP on all datasets. We further explore the related image-importance task in event recognition, demonstrating how the learned attentions correlate with the human-annotated importance for this subjective task, thus opening the door for new applications.

ImageNet-1K serves as the primary dataset for pretraining deep learning models for computer vision tasks. ImageNet-21K dataset, which is bigger and more diverse, is used less frequently for pretraining, mainly due to its complexity, low accessibility, and underestimation of its added value. This paper aims to close this gap, and make high-quality efficient pretraining on ImageNet-21K available for everyone. Via a dedicated preprocessing stage, utilization of WordNet hierarchical structure, and a novel training scheme called semantic softmax, we show that various models significantly benefit from ImageNet-21K pretraining on numerous datasets and tasks, including small mobile-oriented models. We also show that we outperform previous ImageNet-21K pretraining schemes for prominent new models like ViT and Mixer. Our proposed pretraining pipeline is efficient, accessible, and leads to SoTA reproducible results, from a publicly available dataset.

Visual illusions may be explained by the likelihood of patches in real-world images, as argued by input-driven paradigms in Neuro-Science. However, neither the data nor the tools existed in the past to extensively support these explanations. The era of big data opens a new opportunity to study input-driven approaches. We introduce a tool that computes the likelihood of patches, given a large dataset to learn from. Given this tool, we present a model that supports the approach and explains lightness and color visual illusions in a unified manner. Furthermore, our model generates visual illusions in natural images, by applying the same tool, reversely.

Most attempts to represent 3D shapes for deep learning have focused on volumetric grids, multi-view images and point clouds. In this paper we look at the most popular representation of 3D shapes in computer graphics – a triangular mesh – and ask how it can be utilized within deep learning. The few attempts to answer this question propose to adapt convolutions & pooling to suit Convolutional Neural Networks (CNNs). This paper proposes a very different approach, termed MeshWalker to learn the shape directly from a given mesh. The key idea is to represent the mesh by random walks along the surface, which “explore” the meshג€™s geometry and topology. Each walk is organized as a list of vertices, which in some manner imposes regularity on the mesh. The walk is fed into a Recurrent Neural Network (RNN) that “remembers” the history of the walk. We show that our approach achieves state-of-the-art results for two fundamental shape analysis tasks: shape classification and semantic segmentation. Furthermore, even a very small number of examples suffices for learning. This is highly important, since large datasets of meshes are difficult to acquire.

Jigsaw puzzle solving is an intriguing problem which has been explored in computer vision for decades. This paper focuses on a specific variant of the problem – solving puzzles with eroded boundaries. Such erosion makes the problem extremely difficult, since most existing solvers utilize solely the information at the boundaries. Nevertheless, this variant is important since erosion and missing data often occur at the boundaries. The key idea of our proposed approach is to inpaint the eroded boundaries between puzzle pieces and later leverage the quality of the inpainted area to classify a pair of pieces as ‘neighbors or not’. An interesting feature of our architecture is that the same GAN discriminator is used for both inpainting and classification; Training of the second task is simply a continuation of the training of the first, beginning from the point it left off. We show that our approach outperforms other SOTA methods

We propose a new method for anomaly detection of human actions. Our method works directly on human pose graphs that can be computed from an input video sequence. This makes the analysis independent of nuisance parameters such as viewpoint or illumination. We map these graphs to a latent space and cluster them. Each action is then represented by its soft-assignment to each of the clusters. This gives a kind of ”bag of words” representation to the data, where every action is represented by its similarity to a group of base action-words. Then, we use a Dirichlet process based mixture, that is useful for handling proportional data such as our soft-assignment vectors, to determine if an action is normal or not. We evaluate our method on two types of data sets. The first is a fine-grained anomaly detection data set (e.g. ShanghaiTech) where we wish to detect unusual variations of some action. The second is a coarse-grained anomaly detection data set (e.g., a Kinetics-based data set) where few actions are considered normal, and every other action should be considered abnormal. Extensive experiments on the benchmarks show that our method performs considerably better than other state of the art methods.

One of the key ingredients for successful optimization of modern CNNs is identifying a suitable objective. To date, the objective is fixed a-priori at training time, and any variation to it requires re-training a new network. In this paper we present a first attempt at alleviating the need for re-training. Rather than fixing the network at training time, we train a “Dynamic-Net” that can be modified at inference time. Our approach considers an “objective-space” as the space of all linear combinations of two objectives, and the Dynamic-Net is emulating the traversing of this objective-space at test-time, without any further training. We show that this upgrades pre-trained networks by providing an out-of-learning extension, while maintaining the performance quality. The solution we propose is fast and allows a user to interactively modify the network, in real-time, in order to obtain the result he/she desires. We show the benefits of such an approach via several different applications.

This paper deals with the prediction of the memorability of a given image. We start by proposing an algorithm that reaches human-level performance on the LaMem dataset—the only large scale benchmark for memorability prediction. The suggested algorithm is based on three observations we make regarding convolutional neural networks (CNNs) that affect memorability prediction. Having reached human-level performance we were humbled, and asked ourselves whether indeed we have resolved memorability prediction—and answered this question in the negative. We studied a few factors and made some recommendations that should be taken into account when designing the next benchmark.

Large-scale multi-label classification datasets are commonly, and perhaps inevitably, partially annotated. That is, only a small subset of labels are annotated per sample. Different methods for handling the missing labels induce different properties on the model and impact its accuracy. In this work, we analyze the partial labeling problem, then propose a solution based on two key ideas. First, un-annotated labels should be treated selectively according to two probability quantities: the class distribution in the overall dataset and the specific label likelihood for a given data sample. We propose to estimate the class distribution using a dedicated temporary model, and we show its improved efficiency over a naive estimation computed using the dataset’s partial annotations. Second, during the training of the target model, we emphasize the contribution of annotated labels over originally un-annotated labels by using a dedicated asymmetric loss. With our novel approach, we achieve state-of-the-art results on OpenImages dataset (e.g. reaching 87.3 mAP on V6). In addition, experiments conducted on LVIS and simulated-COCO demonstrate the effectiveness of our approach.

This paper re-visits classical problems in document enhancement. Rather than proposing a new algorithm for a specific problem, we introduce a novel general approach. The key idea is to modify any state-of-the-art algorithm, by providing it with new information (input), improving its own results. Interestingly, this information is based on a solution to a seemingly unrelated problem of visibility detection in R3. We show that a simple representation of an image as a 3D point cloud, gives visibility detection on this cloud a new interpretation. What does it mean for a point to be visible? Although this question has been widely studied within computer vision, it has always been assumed that the point set is a sampling of a real scene. We show that the answer to this question in our context reveals unique and useful information about the image. We demonstrate the benefit of this idea for document binarization and for unshadowing.

Maintaining natural image statistics is a crucial factor in restoration and generation of realistic looking images. When training CNNs, photorealism is usually attempted by adversarial training (GAN), that pushes the output images to lie on the manifold of natural images. GANs are very powerful, but not perfect. They are hard to train and the results still often suffer from artifacts. In this paper we propose a complementary approach, that could be applied with or without GAN, whose goal is to train a feed-forward CNN to maintain natural internal statistics. We look explicitly at the distribution of features in an image and train the network to generate images with natural feature distributions. Our approach reduces by orders of magnitude the number of images required for training and achieves state-of-the-art results on both single-image super-resolution, and high-resolution surface normal estimation.

Mesh representation by random walks has been shown to benefit deep learning. Randomness is indeed a powerful concept. However, it comes with a price—some walks might wander around non-characteristic regions of the mesh, which might be harmful to shape analysis, especially when only a few walks are utilized. We propose a novel walk-attention mechanism that leverages the fact that multiple walks are used for a single mesh representation.
The key idea is that the walks may provide each other with information regarding the meaningful (attentive) features of the mesh. We utilize this mutual information to extract a single descriptor of the mesh. This differs from common attention mechanisms that use attention to improve the representation of each individual descriptor. Our approach achieves SOTA results for two basic 3D shape analysis tasks: classification and retrieval. Even a handful of walks along a mesh suffice for learning. Furthermore, our approach provides insight into mesh importance detection.

Puzzle solving is a difficult problem in its own right, even when the pieces are all square and build up a natural image. But what if these ideal conditions do not hold? One such application domain is archaeology, where restoring an artifact from its fragments is highly important. From the point of view of computer vision, archaeological puzzle solving is very challenging, due to three additional difficulties: the fragments are of general shape; they are abraded, especially at the boundaries (where the strongest cues for matching should exist); and the domain of valid transformations between the pieces is continuous. The key contribution of this paper is a fully-automatic and general algorithm that addresses puzzle solving in this intriguing domain. We show that our state-of-the-art approach manages to correctly reassemble dozens of broken artifacts and frescoes.

Maintaining natural image statistics is a crucial factor in restoration and generation of realistic looking images. When training CNNs, photorealism is usually attempted by adversarial training (GAN), that pushes the output images to lie on the manifold of natural images. GANs are very powerful, but not perfect. They are hard to train and the results still often suffer from artifacts. In this paper we propose a complementary approach, that could be applied with or without GAN, whose goal is to train a feed-forward CNN to maintain natural internal statistics. We look explicitly at the distribution of features in an image and train the network to generate images with natural feature distributions. Our approach reduces by orders of magnitude the number of images required for training and achieves state-of-the-art results on both single-image super-resolution, and high-resolution surface normal estimation.

We propose a new type of saliency – context-aware saliency – which aims at detecting the image regions that represent the scene. This definition differs from previous definitions whose goal is to either identify fixation points or detect the dominant object. In accordance with our saliency definition, we present a detection algorithm which is based on four principles observed in the psychological literature.
The benefits of the proposed approach are evaluated in two applications where the context of the dominant objects is just as essential as the objects themselves. In image retargeting we demonstrate that using our saliency prevents distortions in the important regions. In summarization we show that our saliency helps to produce compact, appealing, and informative summaries.

Mesh representation by random walks has been shown to benefit deep learning. Randomness is indeed a powerful concept. However, it comes with a price—some walks might wander around non-characteristic regions of the mesh, which might be harmful to shape analysis, especially when only a few walks are utilized. We propose a novel walk-attention mechanism that leverages the fact that multiple walks are used for a single mesh representation.
The key idea is that the walks may provide each other with information regarding the meaningful (attentive) features of the mesh. We utilize this mutual information to extract a single descriptor of the mesh. This differs from common attention mechanisms that use attention to improve the representation of each individual descriptor. Our approach achieves SOTA results for two basic 3D shape analysis tasks: classification and retrieval. Even a handful of walks along a mesh suffice for learning. Furthermore, our approach provides insight into mesh importance detection.

This paper re-visits classical problems in document enhancement. Rather than proposing a new algorithm for a specific problem, we introduce a novel general approach. The key idea is to modify any state-of-the-art algorithm, by providing it with new information (input), improving its own results. Interestingly, this information is based on a solution to a seemingly unrelated problem of visibility detection in R3. We show that a simple representation of an image as a 3D point cloud, gives visibility detection on this cloud a new interpretation. What does it mean for a point to be visible? Although this question has been widely studied within computer vision, it has always been assumed that the point set is a sampling of a real scene. We show that the answer to this question in our context reveals unique and useful information about the image. We demonstrate the benefit of this idea for document binarization and for unshadowing.

This paper deals with the prediction of the memorability of a given image. We start by proposing an algorithm that reaches human-level performance on the LaMem dataset—the only large scale benchmark for memorability prediction. The suggested algorithm is based on three observations we make regarding convolutional neural networks (CNNs) that affect memorability prediction. Having reached human-level performance we were humbled, and asked ourselves whether indeed we have resolved memorability prediction—and answered this question in the negative. We studied a few factors and made some recommendations that should be taken into account when designing the next benchmark.

One of the key ingredients for successful optimization of modern CNNs is identifying a suitable objective. To date, the objective is fixed a-priori at training time, and any variation to it requires re-training a new network. In this paper we present a first attempt at alleviating the need for re-training. Rather than fixing the network at training time, we train a “Dynamic-Net” that can be modified at inference time. Our approach considers an “objective-space” as the space of all linear combinations of two objectives, and the Dynamic-Net is emulating the traversing of this objective-space at test-time, without any further training. We show that this upgrades pre-trained networks by providing an out-of-learning extension, while maintaining the performance quality. The solution we propose is fast and allows a user to interactively modify the network, in real-time, in order to obtain the result he/she desires. We show the benefits of such an approach via several different applications.

We propose a new type of saliency – context-aware saliency – which aims at detecting the image regions that represent the scene. This definition differs from previous definitions whose goal is to either identify fixation points or detect the dominant object. In accordance with our saliency definition, we present a detection algorithm which is based on four principles observed in the psychological literature.
The benefits of the proposed approach are evaluated in two applications where the context of the dominant objects is just as essential as the objects themselves. In image retargeting we demonstrate that using our saliency prevents distortions in the important regions. In summarization we show that our saliency helps to produce compact, appealing, and informative summaries.