Mohit Talreja, Joshua Diao, Jim James, Radu Casapu, Tejas Santanam, Ethan Mendes, Alan Ritter, Wei Xu, James Hays. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

In this work, we introduce the 3D Gaussian Point Encoder, an explicit per-point embedding built on mixtures of learned 3D Gaussians. This explicit geometric representation for 3D recognition tasks is a departure from widely used implicit representations such as PointNet. However, it is difficult to learn 3D Gaussian encoders in end-to-end fashion with standard optimizers. We develop optimization techniques based on natural gradients and distillation from PointNets to find a Gaussian Basis that can reconstruct PointNet activations. The resulting 3D Gaussian Point Encoders are faster and more parameter efficient than traditional PointNets. As in the 3D reconstruction literature where there has been considerable interest in the move from implicit (e.g., NeRF) to explicit (e.g., Gaussian Splatting) representations, we can take advantage of computational geometry heuristics to accelerate 3D Gaussian Point Encoders further. We extend filtering techniques from 3D Gaussian Splatting to construct encoders that run 2.7× faster as a comparable accuracy PointNet while using 46% less memory and 88% fewer FLOPs. Furthermore, we demonstrate the effectiveness of 3D Gaussian Point Encoders as a component in Mamba3D, running 1.27× faster and achieving a reduction in memory and FLOPs by 42% and 54% respectively. 3D Gaussian Point Encoders are lightweight enough to achieve high framerates on CPU-only devices.

Can a model distinguish between the sound of a spoon hitting a hardwood floor versus a carpeted one? Everyday object interactions produce sounds unique to the objects involved. We introduce the sounding object detection task to evaluate a model's ability to link these sounds to the objects directly involved. Inspired by human perception, our multimodal object-aware framework learns from in-the-wild egocentric videos. To encourage an object-centric approach, we first develop an automatic pipeline to compute segmentation masks of the objects involved to guide the model's focus during training towards the most informative regions of the interaction. A slot attention visual encoder is used to further enforce an object prior. We demonstrate state of the art performance on our new task along with existing multimodal action understanding tasks.

We present personalized residuals and localized attention-guided sampling for efficient concept-driven generation using text-to-image diffusion models. Our method first represents concepts by freezing the weights of a pretrained text-conditioned diffusion model and learning low-rank residuals for a small subset of the model's layers. The residual-based approach then directly enables application of our proposed sampling technique, which applies the learned residuals only in areas where the concept is localized via cross-attention and applies the original diffusion weights in all other regions. Localized sampling therefore combines the learned identity of the concept with the existing generative prior of the underlying diffusion model. We show that personalized residuals effectively capture the identity of a concept in$\sim$3 minutes on a single GPU without the use of regularization images and with fewer parameters than previous models, and localized sampling allows using the original model as strong prior for large parts of the image.

Vision Language Models (VLMs) are rapidly advancing in their capability to answer information-seeking questions. As these models are widely deployed in consumer applications, they could lead to new privacy risks due to emergent abilities to identify people in photos, geolocate images, etc. As we demonstrate, somewhat surprisingly, current open-source and proprietary VLMs are very capable image geolocators, making widespread geolocation with VLMs an immediate privacy risk, rather than merely a theoretical future concern. As a first step to address this challenge, we develop a new benchmark, GPTGeoChat, to test the capability of VLMs to moderate geolocation dialogues with users. We collect a set of 1,000 image geolocation conversations between in-house annotators and GPT-4v, which are annotated with the granularity of location information revealed at each turn. Using this new dataset we evaluate the ability of various VLMs to moderate GPT-4v geolocation conversations by determining when too much location information has been revealed. We find that custom fine-tuned models perform on par with prompted API-based models when identifying leaked location information at the country or city level, however fine-tuning on supervised data appears to be needed to accurately moderate finer granularities, such as the name of a restaurant or building.

Scene flow estimation is the task of describing the 3D motion field between temporally successive point clouds. State-of-the-art methods use strong priors and test-time optimization techniques, but require on the order of tens of seconds to process full-size point clouds, making them unusable as computer vision primitives for real-time applications such as open world object detection. Feedforward methods are considerably faster, running on the order of tens to hundreds of milliseconds for full-size point clouds, but require expensive human supervision. To address both limitations, we propose _Scene Flow via Distillation_, a simple, scalable distillation framework that uses a label-free optimization method to produce pseudo-labels to supervise a feedforward model. Our instantiation of this framework, _ZeroFlow_, achieves **state-of-the-art** performance on the _Argoverse 2 Self-Supervised Scene Flow Challenge_ while using zero human labels by simply training on large-scale, diverse unlabeled data. At test-time, ZeroFlow is over 1000$\times$ faster than label-free state-of-the-art optimization-based methods on full-size point clouds (34 FPS vs 0.028 FPS) and over 1000$\times$ cheaper to train on unlabeled data compared to the cost of human annotation (\\$394 vs ~\\$750,000). To facilitate further research, we will release our code, trained model weights, and high quality pseudo-labels for the Argoverse 2 and Waymo Open datasets.

How easy is it to sneak up on a robot? We examine whether we can detect people using only the incidental sounds they produce as they move, even when they try to be quiet. To do so, we first collect a robotic dataset of high-quality 4-channel audio paired with 360° RGB data of people moving in different indoor settings. Using this dataset, we train models to predict if there is a moving person nearby and then their location using only audio. We implement our method on a robot, allowing it to track a single person moving quietly using only passive audio sensing. For demonstration videos, see our project page.

Touch plays a fundamental role in manipulation for humans; however, machine perception of contact and pressure typically requires invasive sensors. Recent research has shown that deep models can estimate hand pressure based on a single RGB image. However, evaluations have been limited to controlled settings since collecting diverse data with ground-truth pressure measurements is difficult. We present a novel approach that enables diverse data to be captured with only an RGB camera and a cooperative participant. Our key insight is that people can be prompted to apply pressure in a certain way, and this prompt can serve as a weak label to supervise models to perform well under varied conditions. We collect a novel dataset with 51 participants making fingertip contact with diverse objects. Our network, PressureVision++, outperforms human annotators and prior work. We also demonstrate an application of PressureVision++ to mixed reality where pressure estimation allows everyday surfaces to be used as arbitrary touch-sensitive interfaces. Code, data, and models are available online .1

Modern neural networks are over-parameterized and thus rely on strong regularization such as data augmentation and weight decay to reduce overfitting and improve generalization. The dominant form of data augmentation applies invariant transforms, where the learning target of a sample is invariant to the transform applied to that sample. We draw inspiration from human visual classification studies and propose generalizing augmentation with invariant transforms to soft augmentation where the learning target softens non-linearly as a function of the degree of the transform applied to the sample: e.g., more aggressive image crop augmentations produce less confident learning targets. We demonstrate that soft targets allow for more aggressive data augmentation, offer more robust performance boosts, work with other augmentation policies, and interestingly, produce better calibrated models (since they are trained to be less confident on aggressively cropped/occluded examples). Combined with existing aggressive augmentation strategies, soft targets 1) double the top-1 accuracy boost across Cifar-10, Cifar-100, ImageNet-1K, and ImageNet-V2, 2) improve model occlusion performance by up to 4×, and 3) half the expected calibration error (ECE). Finally, we show that soft augmentation generalizes to self-supervised classification tasks. Code available at https://github.com/youngleox/soft_augmentation

Resource-constrained perception systems such as edge computing and vision-for-robotics require vision models to be both accurate and lightweight in computation and memory usage. While knowledge distillation is a proven strategy to enhance the performance of lightweight classification models, its application to structured outputs like object detection and instance segmentation remains a complicated task, due to the variability in outputs and complex internal network modules involved in the distillation process. In this paper, we propose a simple yet surprisingly effective sequential approach to knowledge distillation that progressively transfers the knowledge of a set of teacher detectors to a given lightweight student. To distill knowledge from a highly accurate but complex teacher model, we construct a sequence of teachers to help the student gradually adapt. Our progressive strategy can be easily combined with existing detection distillation mechanisms to consistently maximize student performance in various settings. To the best of our knowledge, we are the first to successfully distill knowledge from Transformer-based teacher detectors to convolution-based students, and unprecedentedly boost the performance of ResNet-50 based RetinaNet from 36.5% to 42.0% AP and Mask R-CNN from 38.2% to 42.5% AP on the MS COCO benchmark. Code available at https://github.com/Shengcao-Cao/MTPD.

State-of-the-art lidar panoptic segmentation (LPS) methods follow “bottom-up” segmentation-centric fashion wherein they build upon semantic segmentation networks by utilizing clustering to obtain object instances. In this paper, we re-think this approach and propose a surprisingly simple yet effective detection-centric network for both LPS and tracking. Our network is modular by design and optimized for all aspects of both the panoptic segmentation and tracking task. One of the core components of our network is the object instance detection branch, which we train using point-level (modal) annotations, as available in segmentation-centric datasets. In the absence of amodal (cuboid) annotations, we regress modal centroids and object extent using trajectory-level supervision that provides information about object size, which cannot be inferred from single scans due to occlusions and the sparse nature of the lidar data. We obtain fine-grained instance segments by learning to associate lidar points with detected centroids. We evaluate our method on several 3D/4D LPS benchmarks and observe that our model establishes a new state-of-the-art among open-sourced models, outperforming recent query-based models.

We focus on the task of far-field 3D detection (Far3Det) of objects beyond a certain distance from an observer, e.g., >50m. Far3Det is particularly important for autonomous vehicles (AVs) operating at highway speeds, which require detections of far-field obstacles to ensure sufficient braking distances. However, contemporary AV benchmarks such as nuScenes underemphasize this problem because they evaluate performance only up to a certain distance (50m). One reason is that obtaining far-field 3D annotations is difficult, particularly for lidar sensors that produce very few point returns for far-away objects. Indeed, we find that almost 50% of far-field objects (beyond 50m) contain zero lidar points. Secondly, current metrics for 3D detection employ a "one-size-fits-all" philosophy, using the same tolerance thresholds for near and far objects, inconsistent with tolerances for both human vision and stereo disparities. Both factors lead to an incomplete analysis of the Far3Det task. For example, while conventional wisdom tells us that highresolution RGB sensors should be vital for 3D detection of far-away objects, lidar-based methods still rank higher compared to RGB counterparts on the current benchmark leaderboards. As a first step towards a Far3Det benchmark, we develop a method to find well-annotated scenes from the nuScenes dataset and derive a well-annotated far-field validation set. We also propose a Far3Det evaluation protocol and explore various 3D detection methods for Far3Det. Our result convincingly justifies the long-held conventional wisdom that high-resolution RGB improves 3D detection in the far-field. We further propose a simple yet effective method that fuses detections from RGB and lidar detectors based on non-maximum suppression, which remarkably outperforms state-of-the-art 3D detectors in the far-field.

We present MSeg, a composite dataset that unifies semantic segmentation datasets from different domains. A naive merge of the constituent datasets yields poor performance due to inconsistent taxonomies and annotation practices. We reconcile the taxonomies and bring the pixel-level annotations into alignment by relabeling more than 220,000 object masks in more than 80,000 images, requiring more than 1.34 years of collective annotator effort. The resulting composite dataset enables training a single semantic segmentation model that functions effectively across domains and generalizes to datasets that were not seen during training. We adopt zero-shot cross-dataset transfer as a benchmark to systematically evaluate a model's robustness and show that MSeg training yields substantially more robust models in comparison to training on individual datasets or naive mixing of datasets without the presented contributions. A model trained on MSeg ranks first on the WildDash-v1 leaderboard for robust semantic segmentation, with no exposure to WildDash data during training. We evaluate our models in the 2020 Robust Vision Challenge (RVC) as an extreme generalization experiment. MSeg training sets include only three of the seven datasets in the RVC; more importantly, the evaluation taxonomy of RVC is different and more detailed. Surprisingly, our model shows competitive performance and ranks second. To evaluate how close we are to the grand aim of robust, efficient, and complete scene understanding, we go beyond semantic segmentation by training instance segmentation and panoptic segmentation models using our dataset. Moreover, we also evaluate various engineering design decisions and metrics, including resolution and computational efficiency. Although our models are far from this grand aim, our comprehensive evaluation is crucial for progress. We share all the models and code with the community.

Soft robotic grippers facilitate contact-rich manipulation, including robust grasping of varied objects. Yet the beneficial compliance of a soft gripper also results in significant deformation that can make precision manipulation challenging. We present visual pressure estimation & control (VPEC), a method that infers pressure applied by a soft gripper using an RGB image from an external camera. We provide results for visual pressure inference when a pneumatic gripper and a tendon-actuated gripper make contact with a flat surface. We also show that VPEC enables precision manipulation via closed-loop control of inferred pressure images. In our evaluation, a mobile manipulator (Stretch RE1 from Hello Robot) uses visual servoing to make contact at a desired pressure; follow a spatial pressure trajectory; and grasp small low-profile objects, including a microSD card, a penny, and a pill. Overall, our results show that visual estimates of applied pressure can enable a soft gripper to perform precision manipulation.

The combination of traditional rendering with neural networks in Deferred Neural Rendering (DNR) [38] provides a compelling balance between computational complexity and realism of the resulting images. Using skinned meshes for rendering articulating objects is a natural extension for the DNR framework and would open it up to a plethora of applications. However, in this case the neural shading step must account for deformations that are possibly not captured in the mesh, as well as alignment in-accuracies and dynamics—which can confound the DNR pipeline. We present Articulated Neural Rendering (ANR), a novel framework based on DNR which explicitly addresses its limitations for virtual human avatars. We show the superiority of ANR not only with respect to DNR but also with methods specialized for avatar creation and animation. In two user studies, we observe a clear preference for our avatar model and we demonstrate state-of-the-art performance on quantitative evaluation metrics. Perceptually, we observe better temporal stability, level of detail and plausibility. More results are available at our project page: https://anr-avatars.github.io.

Acquisition and rendering of photo-realistic human heads is a highly challenging research problem of particular importance for virtual telepresence. Currently, the highest quality is achieved by volumetric approaches trained in a person-specific manner on multi-view data. These models better represent fine structure, such as hair, compared to simpler mesh-based models. Volumetric models typically employ a global code to represent facial expressions, such that they can be driven by a small set of animation parameters. While such architectures achieve impressive rendering quality, they can not easily be extended to the multi-identity setting. In this paper, we devise a novel approach for predicting volumetric avatars of the human head given just a small number of inputs. We enable generalization across identities by a novel parameterization that combines neural radiance fields with local, pixel-aligned features extracted directly from the inputs, thus side-stepping the need for very deep or complex networks. Our approach is trained in an end-to-end manner solely based on a photometric re-rendering loss without requiring explicit 3D supervision. We demonstrate that our approach outperforms the existing state of the art in terms of quality and is able to generate faithful facial expressions in a multi-identity setting.

Scene flow is the three-dimensional (3D) motion field of a scene. It provides information about the spatial arrangement and rate of change of objects in dynamic environments. Current learning-based approaches seek to estimate the scene flow directly from point clouds and have achieved state-of-the-art performance. However, supervised learning methods are inherently domain specific and require a large amount of labeled data. Annotation of scene flow on real-world point clouds is expensive and challenging, and the lack of such datasets has recently sparked interest in self-supervised learning methods. How to accurately and robustly learn scene flow representations without labeled real-world data is still an open problem. Here we present a simple and interpretable objective function to recover the scene flow from point clouds. We use the graph Laplacian of a point cloud to regularize the scene flow to be “asrigid-as-possible”. Our proposed objective function can be used with or without learning-as a self-supervisory signal to learn scene flow representations, or as a non-learning-based method in which the scene flow is optimized during runtime. Our approach outperforms related works in many datasets. We also show the immediate applications of our proposed method for two applications: motion segmentation and point cloud densification.

We present MSeg, a composite dataset that unifies se- mantic segmentation datasets from different domains. A naive merge of the constituent datasets yields poor performance due to inconsistent taxonomies and annotation practices. We reconcile the taxonomies and bring the pixel-level annotations into alignment by relabeling more than 220,000 object masks in more than 80,000 images. The resulting composite dataset enables training a single semantic segmentation model that functions effectively across domains and generalizes to datasets that were not seen during training. We adopt zero-shot cross-dataset transfer as a benchmark to systematically evaluate a model's robustness and show that MSeg training yields substantially more robust models in comparison to training on individual datasets or naive mixing of datasets without the presented contributions. A model trained on MSeg ranks first on the WildDash leaderboard for robust semantic segmentation, with no exposure to WildDash data during training.

Humans involuntarily move their eyes when retrieving an image from memory. This motion is often similar to actually observing the image. We suggest to exploit this behavior as a new modality in human computer interaction, using the motion of the eyes as a descriptor of the image. Interaction requires the user's eyes to be tracked but no voluntary physical activity. We perform a controlled experiment and develop matching techniques using machine learning to investigate if images can be discriminated based on the gaze patterns recorded while users merely think about image. Our results indicate that image retrieval is possible with an accuracy significantly above chance. We also show that this result generalizes to images not used during training of the classifier and extends to uncontrolled settings in a realistic scenario.

Grasping and manipulating objects is an important human skill. Since hand-object contact is fundamental to grasping, capturing it can lead to important insights. However, observing contact through external sensors is challenging because of occlusion and the complexity of the human hand. We present ContactDB, a novel dataset of contact maps for household objects that captures the rich hand-object contact that occurs during grasping, enabled by use of a thermal camera. Participants in our study grasped 3D printed objects with a post-grasp functional intent. ContactDB includes 3750 3D meshes of 50 household objects textured with contact maps and 375K frames of synchronized RGB-D+thermal images. To the best of our knowledge, this is the first large-scale dataset that records detailed contact maps for human grasps. Analysis of this data shows the influence of functional intent and object size on grasping, the tendency to touch/avoid `active areas', and the high frequency of palm and proximal finger contact. Finally, we train state-of-the art image translation and 3D convolution algorithms to predict diverse contact patterns from object shape. Data, code and models are available at https://contactdb.cc.gatech.edu.

We present Argoverse, a dataset designed to support autonomous vehicle perception tasks including 3D tracking and motion forecasting. Argoverse includes sensor data collected by a fleet of autonomous vehicles in Pittsburgh and Miami as well as 3D tracking annotations, 300k extracted interesting vehicle trajectories, and rich semantic maps. The sensor data consists of 360 degree images from 7 cameras with overlapping fields of view, forward-facing stereo imagery, 3D point clouds from long range LiDAR, and 6-DOF pose. Our 290km of mapped lanes contain rich geometric and semantic metadata which are not currently available in any public dataset. All data is released under a Creative Commons license at Argoverse.org. In baseline experiments, we use map information such as lane direction, driveable area, and ground height to improve the accuracy of 3D object tracking. We use 3D object tracking to mine for more than 300k interesting vehicle trajectories to create a trajectory forecasting benchmark. Motion forecasting experiments ranging in complexity from classical methods (k-NN) to LSTMs demonstrate that using detailed vector maps with lane-level information substantially reduces prediction error. Our tracking and forecasting experiments represent only a superficial exploration of the potential of rich maps in robotic perception. We hope that Argoverse will enable the research community to explore these problems in greater depth.

In this paper, we study the task of image retrieval, where the input query is specified in the form of an image plus some text that describes desired modifications to the input image. For example, we may present an image of the Eiffel tower, and ask the system to find images which are visually similar, but are modified in small ways, such as being taken at nighttime instead of during the day. o tackle this task, we embed the query (reference image plus modification text) and the target (images). The encoding function of the image text query learns a representation, such that the similarity with the target image representation is high iff it is a ``positive match''. We propose a new way to combine image and text through residual connection, that is designed for this retrieval task. We show this outperforms existing approaches on 3 different datasets, namely Fashion-200k, MIT-States and a new synthetic dataset we create based on CLEVR. We also show that our approach can be used to perform image classification with compositionally novel labels, and we outperform previous methods on MIT-States on this task.

We propose a novel procedure which adds "content-addressability" to any given unconditional implicit model e.g., a generative adversarial network (GAN). The procedure allows users to control the generative process by specifying a set (arbitrary size) of desired examples based on which similar samples are generated from the model. The proposed approach, based on kernel mean matching, is applicable to any generative models which transform latent vectors to samples, and does not require retraining of the model. Experiments on various high-dimensional image generation problems (CelebA-HQ, LSUN bedroom, bridge, tower) show that our approach is able to generate images which are consistent with the input set, while retaining the image quality of the original model. To our knowledge, this is the first work that attempts to construct, at test time, a content-addressable generative model from a trained marginal model.

Grasping and manipulating objects is an important human skill. Since most objects are designed to be manipulated by human hands, anthropomorphic hands can enable richer human-robot interaction. Desirable grasps are not only stable, but also functional: they enable post-grasp actions with the object. However, functional grasp synthesis for high degree-of-freedom anthropomorphic hands from object shape alone is challenging. We present ContactGrasp, a framework for functional grasp synthesis from object shape and contact on the object surface. Contact can be manually specified or obtained through demonstrations. Our contact representation is object-centric and allows functional grasp synthesis even for hand models different than the one used for demonstration. Using a dataset of contact demonstrations from humans grasping diverse household objects, we synthesize functional grasps for three hand models and two functional intents. The project webpage is https://contactdb.cc.gatech.edu/contactgrasp.html.

This work studies deep metric learning under small to medium scale as we believe that better generalization could be a contributing factor to the improvement of previous fine-grained image retrieval methods; it should be considered when designing future techniques. In particular, we investigate using other layers in a deep metric learning system (besides the embedding layer) for feature extraction and analyze how well they perform on training data and generalize to testing data. From this study, we suggest a new regularization practice where one can add or choose a more optimal layer for feature extraction. State-of-the-art performance is demonstrated on 3 fine-grained image retrieval benchmarks: Cars-196, CUB-200-2011, and Stanford Online Product.

We present a novel computational puzzle solver for square-piece image jigsaw puzzles with no prior information such as piece orientation or anchor pieces. By "piece" we mean a square $d$d x $d$d block of pixels, where we investigate pieces as small as 7 × 7 pixels. To reconstruct such challenging puzzles, we propose to find maximum geometric consensus between pieces, specifically hierarchical piece loops. The proposed algorithm seeks out loops of four pieces and aggregates the smaller loops into higher order "loops of loops" in a bottom-up fashion. In contrast to previous puzzle solvers which aim to maximize compatibility measures between all pairs of pieces and thus depend heavily on the pairwise compatibility measures used, our approach reduces the dependency on the pairwise compatibility measures which become increasingly uninformative for small scales and instead exploits geometric agreement among pieces. Our contribution also includes an improved pairwise compatibility measure which exploits directional derivative information along adjoining boundaries of the pieces. We verify the proposed algorithm as well as its individual components with mathematical analysis and reconstruction experiments.

Maps are a key component in image-based camera localization and visual SLAM systems: they are used to establish geometric constraints between images, correct drift in relative pose estimation, and relocalize cameras after lost tracking. The exact definitions of maps, however, are often application-specific and hand-crafted for different scenarios (e.g. 3D landmarks, lines, planes, bags of visual words). We propose to represent maps as a deep neural net called MapNet, which enables learning a data-driven map representation. Unlike prior work on learning maps, MapNet exploits cheap and ubiquitous sensory inputs like visual odometry and GPS in addition to images and fuses them together for camera localization. Geometric constraints expressed by these inputs, which have traditionally been used in bundle adjustment or pose-graph optimization, are formulated as loss terms in MapNet training and also used during inference. In addition to directly improving localization accuracy, this allows us to update the MapNet (i.e., maps) in a self-supervised manner using additional unlabeled video sequences from the scene. We also propose a novel parameterization for camera rotation which is better suited for deep-learning based camera pose regression. Experimental results on both the indoor 7-Scenes dataset and the outdoor Oxford RobotCar dataset show significant performance improvement over prior work. The MapNet project webpage is https://goo.gl/mRB3Au.

Synthesizing realistic images from human drawn sketches is a challenging problem in computer graphics and vision. Existing approaches either need exact edge maps, or rely on retrieval of existing photographs. In this work, we propose a novel Generative Adversarial Network (GAN) approach that synthesizes plausible images from 50 categories including motorcycles, horses and couches. We demonstrate a data augmentation technique for sketches which is fully automatic, and we show that the augmented data is helpful to our task. We introduce a new network building block suitable for both the generator and discriminator which improves the information flow by injecting the input image at multiple scales. Compared to state-of-the-art image translation methods, our approach generates more realistic images and achieves significantly higher Inception Scores.

In this paper, we investigate deep image synthesis guided by sketch, color, and texture. Previous image synthesis methods can be controlled by sketch and color strokes but we are the first to examine texture control. We allow a user to place a texture patch on a sketch at arbitrary locations and scales to control the desired output texture. Our generative network learns to synthesize objects consistent with these texture suggestions. To achieve this, we develop a local texture loss in addition to adversarial and content loss to train the generative network. We conduct experiments using sketches generated from real images and textures sampled from a separate texture database and results show that our proposed algorithm is able to generate plausible images that are faithful to user controls. Ablation studies show that our proposed pipeline can generate more realistic images than adapting existing methods directly.

Given two candidate models, and a set of target observations, we address the problem of measuring the relative goodness of fit of the two models. We propose two new statistical tests which are nonparametric, computationally efficient (runtime complexity is linear in the sample size), and interpretable. As a unique advantage, our tests can produce a set of examples (informative features) indicating the regions in the data domain where one model fits significantly better than the other. In a real-world problem of comparing GAN models, the test power of our new test matches that of the state-of-the-art test of relative goodness of fit, while being one order of magnitude faster.

We present DeepNav, a Convolutional Neural Network (CNN) based algorithm for navigating large cities using locally visible street-view images. The DeepNav agent learns to reach its destination quickly by making the correct navigation decisions at intersections. We collect a large-scale dataset of street-view images organized in a graph where nodes are connected by roads. This dataset contains 10 city graphs and a total of more than 1 million street-view images. We propose 3 supervised learning approaches for the navigation task, and show how A* search in the city graph can be used to generate labels for the images. Our annotation process is fully automated using publicly available mapping services, and requires no human input. We evaluate the proposed DeepNav models on 4 held-out cities for navigating to 5 different types of destinations and show that our algorithms outperform previous work that uses hand-crafted features and Support Vector Regression (SVR) [19].

Several recent works have used deep convolutional networks to generate realistic imagery. These methods sidestep the traditional computer graphics rendering pipeline and instead generate imagery at the pixel level by learning from large collections of photos (e.g. faces or bedrooms). However, these methods are of limited utility because it is difficult for a user to control what the network produces. In this paper, we propose a deep adversarial image synthesis architecture that is conditioned on sketched boundaries and sparse color strokes to generate realistic cars, bedrooms, or faces. We demonstrate a sketch based image synthesis system which allows users to scribble over the sketch to indicate preferred color for objects. Our network can then generate convincing images that satisfy both the color and the sketch constraints of user. The network is feed-forward which allows users to see the effect of their edits in real time. We compare to recent work on sketch to image synthesis and show that our approach generates more realistic, diverse, and controllable outputs. The architecture is also effective at user-guided colorization of grayscale images.

Image geolocalization, inferring the geographic location of an image, is a challenging computer vision problem with many potential applications. The recent state-of-the-art approach to this problem is a deep image classification approach in which the world is spatially divided into cells and a deep network is trained to predict the correct cell for a given image. We propose to combine this approach with the original Im2GPS approach in which a query image is matched against a database of geotagged images and the location is inferred from the retrieved set. We estimate the geographic location of a query image by applying kernel density estimation to the locations of its nearest neighbors in the reference database. Interestingly, we find that the best features for our retrieval task are derived from networks trained with classification loss even though we do not use a classification approach at test time. Training with classification loss outperforms several deep feature learning methods (e.g. Siamese networks with contrastive of triplet loss) more typical for retrieval applications. Our simple approach achieves state-of-the-art geolocalization accuracy while also requiring significantly less training data.

We propose to leverage concept-level representations for complex event recognition in photographs given limited training examples. We introduce a novel framework to discover event concept attributes from the web and use that to extract semantic features from images and classify them into social event categories with few training examples. Discovered concepts include a variety of objects, scenes, actions and event sub-types, leading to a discriminative and compact representation for event images. Web images are obtained for each discovered event concept and we use (pretrained) CNN features to train concept classifiers. Extensive experiments on challenging event datasets demonstrate that our proposed method outperforms several baselines using deep CNN features directly in classifying images into events with limited training examples. We also demonstrate that our method achieves the best overall accuracy on a dataset with unseen event categories using a single training example.

We propose a Convolutional Neural Network (CNN) based algorithm - StuffNet - for object detection. In addition to the standard convolutional features trained for region proposal and object detection [33], StuffNet uses convolutional features trained for segmentation of objects and 'stuff' (amorphous categories such as ground and water). Through experiments on Pascal VOC 2010, we show the importance of features learnt from stuff segmentation for improving object detection performance. StuffNet improves performance from 18.8% mAP to 23.9% mAP for small objects. We also devise a method to train StuffNet on datasets that do not have stuff segmentation labels. Through experiments on Pascal VOC 2007 and 2012, we demonstrate the effectiveness of this method and show that StuffNet also significantly improves object detection performance on such datasets.

Image matching is a fundamental problem in Computer Vision. In the context of feature-based matching, SIFT and its variants have long excelled in a wide array of applications. However, for ultra-wide baselines, as in the case of aerial images captured under large camera rotations, the appearance variation goes beyond the reach of SIFT and RANSAC. In this paper we propose a data-driven, deep learning-based approach that sidesteps local correspondence by framing the problem as a classification task. Furthermore, we demonstrate that local correspondences can still be useful. To do so we incorporate an attention mechanism to produce a set of probable matches, which allows us to further increase performance. We train our models on a dataset of urban aerial imagery consisting of 'same' and 'different' pairs, collected for this purpose, and characterize the problem via a human study with annotations from Amazon Mechanical Turk. We demonstrate that our models outperform the state-of-the-art on ultra-wide baseline matching and approach human accuracy.

In this paper, we present a novel computational puzzle solver for square-piece image jigsaw puzzles with no prior information such as piece orientation, anchor pieces or resulting dimension of the puzzle. By "piece" we mean a square dxd block of pixels, where we investigate pieces as small as 7x7 pixels. To reconstruct such challenging puzzles, we aim to search for piece configurations which maximize the size of consensus (i.e. grid or loop) configurations which represent a geometric consensus or agreement among pieces. Pieces are considered for addition to the existing assemblies if these pieces increase the size of the consensus configurations. In contrast to previous puzzle solvers which goal for assemblies maximizing compatibility measures between all pairs of pieces and thus depend heavily on the pairwise compatibility measure used, our new approach reduces the dependency on the pairwise compatibility measures which become increasingly uninformative at small scales and instead exploits geometric agreement among pieces. Our contribution also includes an improved pairwise compatibility measure which exploits directional derivative information along adjoining boundaries of the pieces. For the challenging unknown orientation piece puzzles where the size of pieces is small, we reduce assembly error by up to 75% compared with previous algorithms for standard datasets.

We present the Sketchy database , the first large-scale collection of sketch-photo pairs. We ask crowd workers to sketch particular photographic objects sampled from 125 categories and acquire 75,471 sketches of 12,500 objects. The Sketchy database gives us fine-grained associations between particular photos and sketches, and we use this to train cross-domain convolutional networks which embed sketches and photographs in a common feature space. We use our database as a benchmark for fine-grained retrieval and show that our learned representation significantly outperforms both hand-crafted features as well as deep features trained for sketch or photo classification. Beyond image retrieval, we believe the Sketchy database opens up new opportunities for sketch and image understanding and synthesis.

The recent availability of geo-tagged images and rich geospatial data has inspired a number of algorithms for image based geolocalization. Most approaches predict the location of a query image by matching to ground-level images with known locations (e.g., street-view data). However, most of the Earth does not have ground-level reference photos available. Fortunately, more complete coverage is provided by oblique aerial or “bird's eye” imagery. In this work, we localize a ground-level query image by matching it to a reference database of aerial imagery. We use publicly available data to build a dataset of 78K aligned crossview image pairs. The primary challenge for this task is that traditional computer vision approaches cannot handle the wide baseline and appearance variation of these cross-view pairs. We use our dataset to learn a feature representation in which matching views are near one another and mismatched views are far apart. Our proposed approach, Where-CNN, is inspired by deep learning success in face verification and achieves significant improvements over traditional hand-crafted features and existing deep features learned from other large-scale databases. We show the effectiveness of Where-CNN in finding matches between street view and aerial view imagery and demonstrate the ability of our learned features to generalize to novel locations.

This paper introduces the Tropel system which enables non-technical users to create arbitrary visual detectors without first annotating a training set. Our primary contribution is a crowd active learning pipeline that is seeded with only a single positive example and an unlabeled set of training images. We examine the crowd's ability to train visual detectors given severely limited training themselves. This paper presents a series of experiments that reveal the relationship between worker training, worker consensus and the average precision of detectors trained by crowd-in-the-loop active learning. In order to verify the efficacy of our system, we train detectors for bird species that work nearly as well as those trained on the exhaustively labeled CUB 200 dataset at significantly lower cost and with little effort from the end user. To further illustrate the usefulness of our pipeline, we demonstrate qualitative results on unlabeled datasets containing fashion images and street-level photographs of Paris.

We live in a dynamic visual world where the appearance of scenes changes dramatically from hour to hour or season to season. In this work we study "transient scene attributes" -- high level properties which affect scene appearance, such as "snow", "autumn", "dusk", "fog". We define 40 transient attributes and use crowdsourcing to annotate thousands of images from 101 webcams. We use this "transient attribute database" to train regressors that can predict the presence of attributes in novel images. We demonstrate a photo organization method based on predicted attributes. Finally we propose a high-level image editing method which allows a user to adjust the attributes of a scene, e.g. change a scene to be "snowy" or "sunset". To support attribute manipulation we introduce a novel appearance transfer technique which is simple and fast yet competitive with the state-of-the-art. We show that we can convincingly modify many transient attributes in outdoor scenes.

The recent availability of large amounts of geotagged imagery has inspired a number of data driven solutions to the image geolocalization problem. Existing approaches predict the location of a query image by matching it to a database of georeferenced photographs. While there are many geotagged images available on photo sharing and street view sites, most are clustered around landmarks and urban areas. The vast majority of the Earth's land area has no ground level reference photos available, which limits the applicability of all existing image geolocalization methods. On the other hand, there is no shortage of visual and geographic data that densely covers the Earth - we examine overhead imagery and land cover survey data - but the relationship between this data and ground level query photographs is complex. In this paper, we introduce a cross-view feature translation approach to greatly extend the reach of image geolocalization methods. We can often localize a query even if it has no corresponding ground level images in the database. A key idea is to learn the relationship between ground level appearance and overhead appearance and land cover attributes from sparsely available geotagged ground-level images. We perform experiments over a 1600 km2 region containing a variety of scenes and land cover types. For each query, our algorithm produces a probability density over the region of interest.

We significantly extrapolate the field of view of a photograph by learning from a roughly aligned, wide-angle guide image of the same scene category. Our method can extrapolate typical photos into complete panoramas. The extrapolation problem is formulated in the shift-map image synthesis framework. We analyze the self-similarity of the guide image to generate a set of allowable local transformations and apply them to the input image. Our guided shift-map method reserves to the scene layout of the guide image when extrapolating a photograph. While conventional shift-map methods only support translations, this is not expressive enough to characterize the self-similarity of complex scenes. Therefore we additionally allow image transformations of rotation, scaling and reflection. To handle this increase in complexity, we introduce a hierarchical graph optimization method to choose the optimal transformation at each output pixel. We demonstrate our approach on a variety of indoor, outdoor, natural, and man-made scenes.

Blind image deconvolution, i.e., estimating a blur kernel k and a latent image x from an input blurred image y, is a severely ill-posed problem. In this paper we introduce a new patch-based strategy for kernel estimation in blind deconvolution. Our approach estimates a “trusted” subset of x by imposing a patch prior specifically tailored towards modeling the appearance of image edge and corner primitives. To choose proper patch priors we examine both statistical priors learned from a natural image dataset and a simple patch prior from synthetic structures. Based on the patch priors, we iteratively recover the partial latent image x and the blur kernel k. A comprehensive evaluation shows that our approach achieves state-of-the-art results for uniformly blurred images.

In this paper we present the first large-scale scene attribute database. First, we perform crowd-sourced human studies to find a taxonomy of 102 discriminative attributes. Next, we build the “SUN attribute database” on top of the diverse SUN categorical database. Our attribute database spans more than 700 categories and 14,000 images and has potential for use in high-level scene understanding and fine-grained scene recognition. We use our dataset to train attribute classifiers and evaluate how well these relatively simple classifiers can recognize a variety of attributes related to materials, surface properties, lighting, functions and affordances, and spatial envelope properties.

In this paper, we present a highly data-driven approach to the task of single image super-resolution. Super-resolution is a challenging problem due to its massively under-constrained nature - for any low-resolution input there are numerous high-resolution possibilities. Our key observation is that, even with extremely low-res input images, we can use global scene descriptors and Internet-scale image databases to find similar scenes which provide ideal example textures to constrain the image upsampling problem. We quantitatively show that the statistics of scene matches are more predictive than internal image statistics for the super-resolution task. Finally, we build on recent patch-based texture transfer techniques to hallucinate texture detail and compare our super-resolution with other recent methods.

Humans have used sketching to depict our visual world since prehistoric times. Even today, sketching is possibly the only rendering technique readily available to all humans. This paper is the first large scale exploration of human sketches. We analyze the distribution of non-expert sketches of everyday objects such as 'teapot' or 'car'. We ask humans to sketch objects of a given category and gather 20,000 unique sketches evenly distributed over 250 object categories. With this dataset we perform a perceptual study and find that humans can correctly identify the object category of a sketch 73% of the time. We compare human performance against computational recognition methods. We develop a bag-of-features sketch representation and use multi-class support vector machines, trained on our sketch dataset, to classify sketches. The resulting recognition method is able to identify unknown sketches with 56% accuracy (chance is 0.4%). Based on the computational model, we demonstrate an interactive sketch recognition system. We release the complete crowd-sourced dataset of sketches to the community.

As computer vision datasets grow larger, the community is increasingly relying on crowdsourced annotations to train and test their algorithms. Since the capability of online annotators is considered varied and unpredictable, many strategies have been proposed to “clean ” crowdsourced annotations. However, these strategies typically require more annotations, rather than using the annotation or image content itself. In this paper we propose and evaluate several strategies for automatically estimating the quality of an object annotation. Finally, we show that we can significantly outperform simple baselines by combining multiple image-based annotation assessment strategies. 1

Scene categorization is a fundamental problem in computer vision. However, scene understanding research has been constrained by the limited scope of currently-used databases which do not capture the full variety of scene categories. Whereas standard databases for object categorization contain hundreds of different classes of objects, the largest available dataset of scene categories contains only 15 classes. In this paper we propose the extensive Scene UNderstanding (SUN) database that contains 899 categories and 130,519 images. We use 397 well-sampled categories to evaluate numerous state-of-the-art algorithms for scene recognition and establish new bounds of performance. We measure human scene classification performance on the SUN database and compare this with computational methods. Additionally, we study a finer-grained scene representation to detect scenes embedded inside of larger scenes.

This paper presents an empirical evaluation of the role of context in a contemporary, challenging object detection task - the PASCAL VOC 2008. Previous experiments with context have mostly been done on home-grown datasets, often with non-standard baselines, making it difficult to isolate the contribution of contextual information. In this work, we present our analysis on a standard dataset, using top-performing local appearance detectors as baseline. We evaluate several different sources of context and ways to utilize it. While we employ many contextual cues that have been used before, we also propose a few novel ones including the use of geographic context and a new approach for using object spatial support.

This paper presents a method for estimating geographic location for sequences of time-stamped photographs. A prior distribution over travel describes the likelihood of traveling from one location to another during a given time interval. This distribution is based on a training database of 6 million photographs from Flickr.com. An image likelihood for each location is defined by matching a test photograph against the training database. Inferring location for images in a test sequence is then performed using the Forward-Backward algorithm, and the model can be adapted to individual users as well. Using temporal constraints allows our method to geolocate images without recognizable landmarks, and images with no geographic cues whatsoever. This method achieves a substantial performance improvement over the best-available baseline, and geolocates some users' images with near-perfect accuracy.

Estimating geographic information from an image is an excellent, difficult high-level computer vision problem whose time has come. The emergence of vast amounts of geographically-calibrated image data is a great reason for computer vision to start looking globally - on the scale of the entire planet! In this paper, we propose a simple algorithm for estimating a distribution over geographic locations from a single image using a purely data-driven scene matching approach. For this task, we leverage a dataset of over 6 million GPS-tagged images from the Internet. We represent the estimated image location as a probability distribution over the Earthpsilas surface. We quantitatively evaluate our approach in several geolocation tasks and demonstrate encouraging performance (up to 30 times better than chance). We show that geolocation estimates can provide the basis for numerous other image understanding tasks such as population density estimation, land cover estimation or urban/rural classification.

We introduce a novel image segmentation algorithm that uses translational symmetry as the primary foreground/background separation cue. We investigate the process of identifying and analyzing image regions that present approximate translational symmetry for the purpose of image forground/background separation. In conjunction with texture-based inpainting, understanding the different see-through layers allows us to perform powerful image manipulations such as recovering a mesh-occluded background (as much as 53% occluded area) to achieve the effect of image and photo de-fencing. Our algorithm consists of three distinct phases- (1) automatically finding the skeleton structure of a potential frontal layer (fence) in the form of a deformed lattice, (2) separating foreground/background layers using appearance regularity, and (3) occluded foreground inpainting to reveal a complete, non-occluded image. Each of these three tasks presents its own special computational challenges that are not encountered in previous, general image de-layering or texture inpainting applications.

What can you do with a million images? In this paper we present a new image completion algorithm powered by a huge database of photographs gathered from the Web. The algorithm patches up holes in images by finding similar image regions in the database that are not only seamless but also semantically valid. Our chief insight is that while the space of images is effectively infinite, the space of semantically differentiable scenes is actually not that large. For many image completion tasks we are able to find similar scenes which contain image fragments that will convincingly complete the image. Our algorithm is entirely data-driven, requiring no annotations or labelling by the user. Unlike existing image completion methods, our algorithm can generate a diverse set of results for each input image and we allow users to select among them. We demonstrate the superiority of our algorithm over existing image completion approaches.

Designing tensor fields in the plane and on surfaces is a necessary task in many graphics applications, such as painterly rendering, pen-and-ink sketching of smooth surfaces, and anisotropic remeshing. In this article, we present an interactive design system that allows a user to create a wide variety of symmetric tensor fields over 3D surfaces either from scratch or by modifying a meaningful input tensor field such as the curvature tensor. Our system converts each user specification into a basis tensor field and combines them with the input field to make an initial tensor field. However, such a field often contains unwanted degenerate points which cannot always be eliminated due to topological constraints of the underlying surface. To reduce the artifacts caused by these degenerate points, our system allows the user to move a degenerate point or to cancel a pair of degenerate points that have opposite tensor indices. These operations provide control over the number and location of the degenerate points in the field. We observe that a tensor field can be locally converted into a vector field so that there is a one-to-one correspondence between the set of degenerate points in the tensor field and the set of singularities in the vector field. This conversion allows us to effectively perform degenerate point pair cancellation and movement by using similar operations for vector fields. In addition, we adapt the image-based flow visualization technique to tensor fields, therefore allowing interactive display of tensor fields on surfaces. We demonstrate the capabilities of our tensor field design system with painterly rendering, pen-and-ink sketching of surfaces, and anisotropic remeshing.

Near regular textures are pervasive in man-made and natural world. Their global regularity and local randomness pose new difficulties to the state of the art texture analysis and synthesis algorithms. We carry out a systematic comparison study on the performance of four texture synthesis algorithms on near-regular textures. Our results confirm that faithful near-regular texture synthesis remains a challenging problem for the state of the art general purpose texture synthesis algorithms. In addition, we provide comparison of human perception with computer evaluations on the quality of the texture synthesis results.

A near-regular texture deviates geometrically and photometrically from a regular congruent tiling. Although near-regular textures are ubiquitous in the man-made and natural world, they present computational challenges for state of the art texture analysis and synthesis algorithms. Using regular tiling as our anchor point, and with user-assisted lattice extraction, we can explicitly model the deformation of a near-regular texture with respect to geometry, lighting and color. We treat a deformation field both as a function that acts on a texture and as a texture that is acted upon, and develop a multi-modal framework where each deformation field is subject to analysis, synthesis and manipulation. Using this formalization, we are able to construct simple parametric models to faithfully synthesize the appearance of a near-regular texture and purposefully control its regularity.

WC describe PRMS an animated simulation of process and resource management in a multi-programming operating system.PRMS was developed to illustrate important operating system concepts including process creation, process state transition, context switching, process scheduling, resource scheduling, interrupt handling, and deadlock detection and recovery.The users write their own, programs in a specified format, submit these to the system for execution, and monitor "behind the scenes" mechanisms and the behavior of the system on the screen.In addition to illustrating and reinforcing operating system concepts, PRMS can also be used for course projects by having the students change an algorithm such as the CPU scheduling algorithm or memory management scheme or add new features such as device interrupts to the system.