Topic
Information projection
About: Information projection is a research topic. Over the lifetime, 118 publications have been published within this topic receiving 983 citations.
Papers published on a yearly basis
Papers
TL;DR: This paper presents a novel framework for learning a generative image representation-the hybrid image template (HIT) from a small number of image examples, associated with a well-normalized probability model that integrates the heterogeneous feature statistics.
Abstract: This paper presents a novel framework for learning a generative image representation-the hybrid image template (HIT) from a small number (i.e., 3 \sim 20) of image examples. Each learned template is composed of, typically, 50 \sim 500 image patches whose geometric attributes (location, scale, orientation) may adapt in a local neighborhood for deformation, and whose appearances are characterized, respectively, by four types of descriptors: local sketch (edge or bar), texture gradients with orientations, flatness regions, and colors. These heterogeneous patches are automatically ranked and selected from a large pool according to their information gains using an information projection framework. Intuitively, a patch has a higher information gain if 1) its feature statistics are consistent within the training examples and are distinctive from the statistics of negative examples (i.e., generic images or examples from other categories); and 2) its feature statistics have less intraclass variations. The learning process pursues the most informative (for either generative or discriminative purpose) patches one at a time and stops when the information gain is within statistical fluctuation. The template is associated with a well-normalized probability model that integrates the heterogeneous feature statistics. This automated feature selection procedure allows our algorithm to scale up to a wide range of image categories, from those with regular shapes to those with stochastic texture. The learned representation captures the intrinsic characteristics of the object or scene categories. We evaluate the hybrid image templates on several public benchmarks, and demonstrate classification performances on par with state-of-the-art methods like HoG+SVM, and when small training sample sizes are used, the proposed system shows a clear advantage.
120 citations
TL;DR: Theoretical analysis shows that conventional linear projection methods such as (weighted) PCA, maximum margin criterion (MMC), linear discriminant analysis (LDA), and LPP could be derived from the LPMIP framework by setting different graph models and constraints.
Abstract: Dimensionality reduction is usually involved in the domains of artificial intelligence and machine learning. Linear projection of features is of particular interest for dimensionality reduction since it is simple to calculate and analytically analyze. In this paper, we propose an essentially linear projection technique, called locality-preserved maximum information projection (LPMIP), to identify the underlying manifold structure of a data set. LPMIP considers both the within-locality and the between-locality in the processing of manifold learning. Equivalently, the goal of LPMIP is to preserve the local structure while maximize the out-of-locality (global) information of the samples simultaneously. Different from principal component analysis (PCA) that aims to preserve the global information and locality-preserving projections (LPPs) that is in favor of preserving the local structure of the data set, LPMIP seeks a tradeoff between the global and local structures, which is adjusted by a parameter alpha, so as to find a sub- space that detects the intrinsic manifold structure for classification tasks. Computationally, by constructing the adjacency matrix, LPMIP is formulated as an eigenvalue problem. LPMIP yields orthogonal basis functions, and completely avoids the singularity problem as it exists in LPP. Further, we develop an efficient and stable LPMIP/QR algorithm for implementing LPMIP, especially, on high-dimensional data set. Theoretical analysis shows that conventional linear projection methods such as (weighted) PCA, maximum margin criterion (MMC), linear discriminant analysis (LDA), and LPP could be derived from the LPMIP framework by setting different graph models and constraints. Extensive experiments on face, digit, and facial expression recognition show the effectiveness of the proposed LPMIP method.
91 citations
TL;DR: In this paper, the authors introduce the concept of information projection and provide a simple but widely applicable model, which describes a novel agency con-ict in a frictionless learning environment, where biased evaluators exaggerate how much experts could have known ex-ante and underestimate experts on average.
Abstract: People exaggerate the extent to which their information is shared with others. This paper introduces the concept of such information projection and provides a simple but widely applicable model. The key application describes a novel agency con‡ict in a frictionless learning environment. When monitoring with ex-post information, biased evaluators exaggerate how much experts could have known ex-ante and underestimate experts on average. Experts, to defend their reputations, are too eager to base predictions on ex-ante information which substitutes for the information jurors independently learn ex-post, and too reluctant to base predictions on ex-ante information which complements the information jurors independently learn ex-post. Instruments which mitigate
74 citations
Posted Content•
TL;DR: This paper introduces the concept of such information projection and provides a simple but widely applicable model that describes a novel agency conflict in a frictionless learning environment and applications to defensive medicine are discussed.
Abstract: People exaggerate the extent to which their information is shared with others. This paper introduces the concept of such information projection and provides a simple but widely applicable model. The key application describes a novel agency conflict in a frictionless learning environment. When monitoring with ex post information, biased evaluators exaggerate how much experts could have known ex ante and underestimate experts on average. Experts, to defend their reputations, are too eager to base predictions on ex ante information that substitutes for the information jurors independently learn ex post and too reluctant to base predictions on ex ante information that complements the information jurors independently learn ex post. Instruments that mitigate Bayesian agency conflicts are either ineffective or directly backfire. Limiting monitoring improves efficiency. Applications to defensive medicine are discussed.
60 citations
Patent•
19 Jul 2001TL;DR: In this article, a three-dimensional position information calculation was proposed to enable a user to easily understand the movement of a moving body in an image processing apparatus for making a display device display the pickup image of a driver recorder.
Abstract: PROBLEM TO BE SOLVED: To enable a user to easily understand the movement of a moving body in an image processing apparatus for making a display device display the pickup image of a driver recorder SOLUTION: A three-dimensional position information calculation part 11 calculates a three-dimensional position information from a time-series stereo pickup image from a drive recorder, and a moving body extraction part 12a extracts the same moving body, and a face setting part 13 sets a projection face desired by a user when displaying a pickup image, that is, the direction of a line of sight, and a three-dimensional position information integration part 15 integrates time-series images on the set projection face, and a three-dimensional position information calculation part 16 calculates each position of the moving body on the integrated screen, and a three-dimensional position information projection part 14 makes a display device 3 display it Therefore, it is possible to convert the movement of the moving body analyzed from the time-series three-dimensional pickup image into an image viewed from the line of sight of a driver or the line of sight of the eyewitness of an accident for display Thus, it is possible for a user to easily understand the movement of the moving body COPYRIGHT: (C)2011,JPO&INPIT
46 citations
Related Topics (5)
Performance Metrics
| Year | Papers |
|---|---|
| 2021 | 8 |
| 2020 | 7 |
| 2019 | 3 |
| 2018 | 13 |
| 2017 | 11 |
| 2016 | 16 |