Showing papers by "Antonio Torralba published in 2013"

SUN3D: A Database of Big Spaces Reconstructed Using SfM and Object Labels

Abstract: Existing scene understanding datasets contain only a limited set of views of a place, and they lack representations of complete 3D spaces. In this paper, we introduce SUN3D, a large-scale RGB-D video database with camera pose and object labels, capturing the full 3D extent of many places. The tasks that go into constructing such a dataset are difficult in isolation -- hand-labeling videos is painstaking, and structure from motion (SfM) is unreliable for large spaces. But if we combine them together, we make the dataset construction task much easier. First, we introduce an intuitive labeling tool that uses a partial reconstruction to propagate labels from one frame to another. Then we use the object labels to fix errors in the reconstruction. For this, we introduce a generalization of bundle adjustment that incorporates object-to-object correspondences. This algorithm works by constraining points for the same object from different frames to lie inside a fixed-size bounding box, parameterized by its rotation and translation. The SUN3D database, the source code for the generalized bundle adjustment, and the web-based 3D annotation tool are all available at http://sun3d.cs.princeton.edu.

HOGgles: Visualizing Object Detection Features

Abstract: We introduce algorithms to visualize feature spaces used by object detectors. The tools in this paper allow a human to put on 'HOG goggles' and perceive the visual world as a HOG based object detector sees it. We found that these visualizations allow us to analyze object detection systems in new ways and gain new insight into the detector's failures. For example, when we visualize the features for high scoring false alarms, we discovered that, although they are clearly wrong in image space, they do look deceptively similar to true positives in feature space. This result suggests that many of these false alarms are caused by our choice of feature space, and indicates that creating a better learning algorithm or building bigger datasets is unlikely to correct these errors. By visualizing feature spaces, we can gain a more intuitive understanding of our detection systems.

Parsing IKEA Objects: Fine Pose Estimation

Abstract: We address the problem of localizing and estimating the fine-pose of objects in the image with exact 3D models. Our main focus is to unify contributions from the 1970s with recent advances in object detection: use local keypoint detectors to find candidate poses and score global alignment of each candidate pose to the image. Moreover, we also provide a new dataset containing fine-aligned objects with their exactly matched 3D models, and a set of models for widely used objects. We also evaluate our algorithm both on object detection and fine pose estimation, and show that our method outperforms state-of-the art algorithms.

Learning with Hierarchical-Deep Models

Abstract: We introduce HD (or “Hierarchical-Deep”) models, a new compositional learning architecture that integrates deep learning models with structured hierarchical Bayesian (HB) models. Specifically, we show how we can learn a hierarchical Dirichlet process (HDP) prior over the activities of the top-level features in a deep Boltzmann machine (DBM). This compound HDP-DBM model learns to learn novel concepts from very few training example by learning low-level generic features, high-level features that capture correlations among low-level features, and a category hierarchy for sharing priors over the high-level features that are typical of different kinds of concepts. We present efficient learning and inference algorithms for the HDP-DBM model and show that it is able to learn new concepts from very few examples on CIFAR-100 object recognition, handwritten character recognition, and human motion capture datasets.

Modifying the Memorability of Face Photographs

Abstract: Contemporary life bombards us with many new images of faces every day, which poses non-trivial constraints on human memory. The vast majority of face photographs are intended to be remembered, either because of personal relevance, commercial interests or because the pictures were deliberately designed to be memorable. Can we make a portrait more memorable or more forgettable automatically? Here, we provide a method to modify the memorability of individual face photographs, while keeping the identity and other facial traits (e.g. age, attractiveness, and emotional magnitude) of the individual fixed. We show that face photographs manipulated to be more memorable (or more forgettable) are indeed more often remembered (or forgotten) in a crowd-sourcing experiment with an accuracy of 74%. Quantifying and modifying the 'memorability' of a face lends itself to many useful applications in computer vision and graphics, such as mnemonic aids for learning, photo editing applications for social networks and tools for designing memorable advertisements.

Are all training examples equally valuable

Abstract: When learning a new concept, not all training examples may prove equally useful for training: some may have higher or lower training value than others The goal of this paper is to bring to the attention of the vision community the following considerations: (1) some examples are better than others for training detectors or classifiers, and (2) in the presence of better examples, some examples may negatively impact performance and removing them may be beneficial In this paper, we propose an approach for measuring the training value of an example, and use it for ranking and greedily sorting examples We test our methods on different vision tasks, models, datasets and classifiers Our experiments show that the performance of current state-of-the-art detectors and classifiers can be improved when training on a subset, rather than the whole training set

Basic level scene understanding: categories, attributes and structures

Abstract: A longstanding goal of computer vision is to build a system that can automatically understand a 3D scene from a single image. This requires extracting semantic concepts and 3D information from 2D images which can depict an enormous variety of environments that comprise our visual world. This paper summarizes our recent efforts toward these goals. First, we describe the richly annotated SUN database which is a collection of annotated images spanning 908 different scene categories with object, attribute, and geometric labels for many scenes. This database allows us to systematically study the space of scenes and to establish a benchmark for scene and object recognition. We augment the categorical SUN database with 102 scene attributes for every image and explore attribute recognition. Finally, we present an integrated system to extract the 3D structure of the scene and objects depicted in an image.

Shape Anchors for Data-Driven Multi-view Reconstruction

Abstract: We present a data-driven method for building dense 3D reconstructions using a combination of recognition and multi-view cues. Our approach is based on the idea that there are image patches that are so distinctive that we can accurately estimate their latent 3D shapes solely using recognition. We call these patches shape anchors, and we use them as the basis of a multi-view reconstruction system that transfers dense, complex geometry between scenes. We "anchor" our 3D interpretation from these patches, using them to predict geometry for parts of the scene that are relatively ambiguous. The resulting algorithm produces dense reconstructions from stereo point clouds that are sparse and noisy, and we demonstrate it on a challenging dataset of real-world, indoor scenes.

A 2d + 3d rich data approach to scene understanding

Abstract: On your one-minute walk from the coffee machine to your desk each morning, you pass by dozens of scenes - a kitchen, an elevator, your office - and you effortlessly recognize them and perceive their 3D structure. But this one-minute scene-understanding problem has been an open challenge in computer vision since the field was first established 50 years ago. In this dissertation, we aim to rethink the path researchers took over these years, challenge the standard practices and implicit assumptions in the current research, and redefine several basic principles in computational scene understanding. The key idea of this dissertation is that learning from rich data under natural setting is crucial for finding the right representation for scene understanding. First of all, to overcome the limitations of object-centric datasets, we built the Scene Understanding (SUN) Database, a large collection of real-world images that exhaustively spans all scene categories. This scene-centric dataset provides a more natural sample of human visual world, and establishes a realistic benchmark for standard 2D recognition tasks. However, while an image is a 2D array, the world is 3D and our eyes see it from a viewpoint, but this is not traditionally modeled. To obtain a 3D understanding at high-level, we reintroduce geometric figures using modern machinery. To model scene viewpoint, we propose a panoramic place representation to go beyond aperture computer vision and use data that is close to natural input for human visual system. This paradigm shift toward rich representation also opens up new challenges that require a new kind of big data - data with extra descriptions, namely rich data. Specifically, we focus on a highly valuable kind of rich data - multiple viewpoints in 3D - and we build the SUN3D data base to obtain an integrated place-centric representation of scenes. We argue for the great importance of modeling the computer's role as an agent in a 3D scene, and demonstrate the power of place-centric scene representation. (Copies available exclusively from MIT Libraries, libraries.mit.edu/docs - [email protected])