Open Access. Powered by Scholars. Published by Universities.®
Physical Sciences and Mathematics Commons™
Open Access. Powered by Scholars. Published by Universities.®
Articles 1 - 9 of 9
Full-Text Articles in Physical Sciences and Mathematics
Bounding The Probability Of Error For High Precision Optical Character Recognition, Gary B. Huang, Andrew Kae, Carl Doersch, Erik G. Learned-Miller
Bounding The Probability Of Error For High Precision Optical Character Recognition, Gary B. Huang, Andrew Kae, Carl Doersch, Erik G. Learned-Miller
Erik G Learned-Miller
We consider a model for which it is important, early in proces sing, to estimate some variables with high precision, but perhaps at relatively low recall. I f some variables can be identified with near certainty, they can be conditioned upon, allowing furt her inference to be done efficiently. Specifically, we consider optical character recognition (O CR) systems that can be bootstrapped by identifying a subset of correctly translated document wo rds with very high precision. This “clean set” is subsequently used as document-specific train ing data. While OCR systems produce confidence measures for the identity of each letter or …
Learning To Align From Scratch, Gary Huang, Marwan Mattar, Honglak Lee, Erik Learned-Miller
Learning To Align From Scratch, Gary Huang, Marwan Mattar, Honglak Lee, Erik Learned-Miller
Erik G Learned-Miller
Unsupervised joint alignment of images has been demonstrated to improve performance on recognition tasks such as face verification. Such alignment reduces undesired variability due to factors such as pose, while only requiring weak supervision in the form of poorly aligned examples. However, prior work on unsupervised alignment of complex, real-world images has required the careful selection of feature representation based on hand-crafted image descriptors, in order to achieve an appropriate, smooth optimization landscape. In this paper, we instead propose a novel combination of unsupervised joint alignment with unsupervised feature learning. Specifically, we incorporate deep learning into the congealing alignment framework. …
Distribution Fields For Tracking, Erik Learned-Miller, Laura Lara
Distribution Fields For Tracking, Erik Learned-Miller, Laura Lara
Erik G Learned-Miller
Visual tracking of general objects often relies on the assumption that gradient descent of the alignment function will reach the global optimum. A common technique to smooth the objective function is to blur the image. However, blurring the image destroys image information, which can cause the target to be lost. To address this problem we introduce a method for building an image descriptor using distribution fields (DFs), a representation that allows smoothing the objective function without destroying information about pixel values. We present experimental evidence on the superiority of the width of the basin of attraction around the global optimum …
Background Modeling Using Adaptive Pixelwise Kernel Variances In A Hybrid Feature Space, Erik G. Learned-Miller, Manjunath Narayana, Allen Hanson
Background Modeling Using Adaptive Pixelwise Kernel Variances In A Hybrid Feature Space, Erik G. Learned-Miller, Manjunath Narayana, Allen Hanson
Erik G Learned-Miller
Recent work on background subtraction has shown de- velopments on two major fronts. In one, there has been increasing sophistication of probabilistic models, from mix- tures of Gaussians at each pixel [ 7 ], to kernel density esti- mates at each pixel [ 1 ], and more recently to joint domain- range density estimates that incorporate spatial informa- tion [ 6 ]. Another line of work has shown the benefits of increasingly complex feature representations, including the use of texture information, local binary patterns, and re- cently scale-invariant local ternary patterns [ 4 ]. In this work, we use joint …
Scene Text Recognition With Bilateral Regression, Jacqueline Feild, Erik G. Learned-Miller
Scene Text Recognition With Bilateral Regression, Jacqueline Feild, Erik G. Learned-Miller
Erik G Learned-Miller
This paper focuses on improving the recognition of text in images of natural scenes, such as storefront signs or street signs. This is a difficult problem due to lighting conditions, variation in font shape and color, and complex backgrounds. We present a word recognition system that addresses these difficulties using an innovative technique to extract and recognize foreground text in an image. First, we develop a new method, called bilateral regression, for extracting and modeling one coherent (although not necessarily contiguous) region from an image. The method models smooth color changes across an image region without being corrupted by neighboring …
Unsupervised Joint Alignment And Clustering Using Bayesian Nonparametrics, Marwan Mattar, Allen Hanson, Erik G. Learned-Miller
Unsupervised Joint Alignment And Clustering Using Bayesian Nonparametrics, Marwan Mattar, Allen Hanson, Erik G. Learned-Miller
Erik G Learned-Miller
Joint alignment of a collection of functions is the process of independently transforming the func- tions so that they appear more similar to each other. Typically, such unsupervised alignment al- gorithms fail when presented with complex data sets arising from multiple modalities or make re- strictive assumptions about the form of the func- tions or transformations, limiting their general- ity. We present a transformed Bayesian infinite mixture model that can simultaneously align and cluster a data set. Our model and associated learning scheme offer two key advantages: the optimal number of clusters is determined in a data-driven fashion through the …
Learning Hierarchical Representations For Face Verification, Gary B. Huang, Honglak Lee, Erik G. Learned-Miller
Learning Hierarchical Representations For Face Verification, Gary B. Huang, Honglak Lee, Erik G. Learned-Miller
Erik G Learned-Miller
Most modern face recognition systems rely on a feature representation given by a hand-crafted image descriptor, such as Local Binary Patterns (LBP), and achieve improved performance by combining several such representations. In this paper, we propose deep learning as a natural source for obtaining additional, complementary representations. To learn features in high-resolution images, we make use of convolutional deep belief networks. Moreover, to take advantage of global structure in an object class, we develop local convolutional restricted Boltzmann machines, a novel convolutional learning model that exploits the global structure by not assuming stationarity of features across the image, while maintaining …
Half-Wits: Software Techniques For Low-Voltage Probabilistic Storage On Microcontrollers With Nor Flash Memory., Mastooreh Salajegheh, Yue Wang, Anxiao Jiang, Erik G. Learned-Miller, Kevin Fu
Half-Wits: Software Techniques For Low-Voltage Probabilistic Storage On Microcontrollers With Nor Flash Memory., Mastooreh Salajegheh, Yue Wang, Anxiao Jiang, Erik G. Learned-Miller, Kevin Fu
Erik G Learned-Miller
No abstract provided.
Improvements In Joint Domain-Range Modeling For Background Subtraction, Manjunath Narayana, Allen Hanson, Erik G. Learned-Miller
Improvements In Joint Domain-Range Modeling For Background Subtraction, Manjunath Narayana, Allen Hanson, Erik G. Learned-Miller
Erik G Learned-Miller
In many algorithms for background modeling, a distribution over feature values is modeled at each pixel. These models, however, do not account for the dependencies that may exist among nearby pixels. The joint domain-range kernel density estimate (KDE) model by Sheikh and Shah [ 7 ], which is not a pixel-wise model, represents the background and foreground processes by combining the three color dimensions and two spatial dimensions into a five-dimensional joint space. The Sheikh and Shah model, as we will show, has a peculiar dependence on the size of the image. In contrast, we build three-dimensional color distributions at …