“Energy” models for continuous domains can be applied to many problems, but often suffer from high computational expense in training, due to the need to repeatedly minimize the energy function to high accuracy. This paper considers a modified setting, where the model is trained in terms of results after optimization is truncated to a fixed number of iterations. We derive “backpropagating” versions of gradient descent, heavy-ball and LBFGS. These are simple to use, as they require as input only routines to compute the gradient of the energy with respect to the domain and parameters. Experimental results on denoising and image labeling problems show that learning with truncated optimization greatly reduces computational expense compared to “full” fitting.

/pdf/generic-methods-for-optimization-based-modeling-1fa76e0brc.pdf

Generic Methods for Optimization-Based Modeling

This paper addresses the problem of recognizing shadows from monochromatic natural images. Without chromatic information, shadow classification is very challenging because the invariant color cues are unavailable. Natural scenes make this problem even harder because of ambiguity from many near black objects. We propose to use both shadow-variant and shadow-invariant cues from illumination, textural and odd order derivative characteristics. Such features are used to train a classifier from boosting a decision tree and integrated into a Conditional random Field, which can enforce local consistency over pixel labels. The proposed approach is evaluated using both qualitative and quantitative results based on a novel database of hand-labeled shadows. Our results show shadowed areas of an image can be identified using proposed monochromatic cues.

/pdf/learning-to-recognize-shadows-in-monochromatic-natural-182svu3t15.pdf

Learning to recognize shadows in monochromatic natural images

Apparatus, systems, and methods are provided for substantially continuous biometric identification (CBID) of an individual using eye signals in real time. The apparatus is included within a wearable computing device with identification of the device wearer based on iris recognition within one or more cameras directed at one or both eyes, and/or other physiological, anatomical and/or behavioral measures. Verification of device user identity can be used to enable or disable the display of secure information. Identity verification can also be included within information that is transmitted from the device in order to determine appropriate security measures by remote processing units. The apparatus may be incorporated within wearable computing that performs other functions including vision correction, head-mounted display, viewing the surrounding environment using scene camera(s), recording audio data via a microphone, and/or other sensing equipment.

Systems and methods for discerning eye signals and continuous biometric identification

Bottom-up segmentation based only on low-level cues is a notoriously difficult problem. This difficulty has lead to recent top-down segmentation algorithms that are based on class-specific image information. Despite the success of top-down algorithms, they often give coarse segmentations that can be significantly refined using low-level cues. This raises the question of how to combine both top-down and bottom-up cues in a principled manner. In this paper we approach this problem using supervised learning. Given a training set of ground truth segmentations we train a fragment-based segmentation algorithm which takes into account both bottom-up and top-down cues simultaneously, in contrast to most existing algorithms which train top-down and bottom-up modules separately. We formulate the problem in the framework of Conditional Random Fields (CRF) and derive a novel feature induction algorithm for CRF, which allows us to efficiently search over thousands of candidate fragments. Whereas pure top-down algorithms often require hundreds of fragments, our simultaneous learning procedure yields algorithms with a handful of fragments that are combined with low-level cues to efficiently compute high quality segmentations.

Learning to Combine Bottom-Up and Top-Down Segmentation

In contrast to the existing approaches that use discrete Conditional Random Field (CRF) models, we propose to use a Gaussian CRF model for the task of semantic segmentation. We propose a novel deep network, which we refer to as Gaussian Mean Field (GMF) network, whose layers perform mean field inference over a Gaussian CRF. The proposed GMF network has the desired property that each of its layers produces an output that is closer to the maximum a posteriori solution of the Gaussian CRF compared to its input. By combining the proposed GMF network with deep Convolutional Neural Networks (CNNs), we propose a new end-to-end trainable Gaussian conditional random field network. The proposed Gaussian CRF network is composed of three sub-networks: (i) a CNN-based unary network for generating unary potentials, (ii) a CNN-based pairwise network for generating pairwise potentials, and (iii) a GMF network for performing Gaussian CRF inference. When trained end-to-end in a discriminative fashion, and evaluated on the challenging PASCALVOC 2012 segmentation dataset, the proposed Gaussian CRF network outperforms various recent semantic segmentation approaches that combine CNNs with discrete CRF models.

/pdf/gaussian-conditional-random-field-network-for-semantic-2ly6baknez.pdf

Gaussian Conditional Random Field Network for Semantic Segmentation

A method and system for synthesizing multiple fingerprint images into a single synthesized fingerprint template. Sets of features are extracted from each of three or more fingerprint images. Pair-wise comparisons identifying correspondences between sets of features are performed between each set of features and every other set of features. Transformations (translation and rotation) for each set of features are simultaneously calculated based on the pair-wise correspondences, and each set of features is transformed accordingly. A synthesized fingerprint template is generated by simultaneously registering the transformed sets of features.

A method for n-wise registration and mosaicing of partial prints

A system and method for inpainting areas in a fingerprint image is provided. The method includes the steps of dividing a fingerprint image into a plurality of image blocks (506) and computing a plurality of block scores for the plurality of image blocks (508). The method also includes generating a blur matrix for the fingerprint image based on the plurality of block scores (510). The method further includes deriving an inpaint region (IR) matrix for the fingerprint image based on a weighting function and the blur matrix, the IR matrix identifying a portion of the plurality of image blocks for inpainting (512, 514).

Automatic identification of fingerprint inpainting target areas

Graphical models are fundamental tools for modeling images and other applications. In this paper, we propose the logistic random field (LRF) model for representing a discrete-valued graphical model. The LRF model is based on an underlying quadratic model and a logistic function. The chief advantages of the LRF are its convenience and flexibility. The quadratic model makes inference easy to implement using standard numerical linear algebra routines. This quadratic model also allows the log-likelihood of the training data to be differentiated with respect to any parameter in the model, enhancing the flexibility of the LRF model. To demonstrate the usefulness of this model we use it to learn how to segment objects, specifically roads, horses, and cows. In addition, we demonstrate the flexibility of the LRF model by incorporating super-pixels. We then show that the LRF segmentation model produces segmentations that are competitive with recently published results.

https://www.cs.ucf.edu/~mtappen/pubs/cvpr08.pdf

The Logistic Random Field — A convenient graphical model for learning parameters for MRF-based labeling

Systems and methods for secure communications in a communications network ( 100 ), including a server ( 102 ) and a client device ( 104 a, 104 b ), are provided. A method includes transmitting a communications session request from the client device to the server ( 304 ), the session request including an in-situ biometric code associated with a biometric sample of a user. The method also includes separately comparing at the server and the client device the in-situ biometric code and an associated canonical biometric code based on comparison criterion ( 208, 310 ), the canonical biometric code representing a biometric reference for the user. The method also includes generating an encryption key at the server and at the client device based on the in-situ and canonical biometric codes ( 218, 314 ). The method further includes establishing an encrypted communications session between the server and the client device if the in-situ biometric code meets comparison criteria at the server and at client device ( 220, 316 ).

Secure digital communications via biometric key generation

A method may be for operating a fingerprint matcher receiving reference fingerprint data. The fingerprint matcher may include a memory and a controller cooperating therewith. The method may include determining ridge flow direction magnitude values for each block of input fingerprint data using the memory and controller, and iteratively identifying blocks of the input fingerprint data in which the respective ridge flow direction magnitude values exceed an iteratively decremented threshold until reaching a stopping point thereby defining a final set of identified blocks of the input fingerprint data using the memory and controller. The method may also include determining a match between the reference fingerprint data and the final set of identified blocks of the input fingerprint data using the memory and controller.

David Lyle

Papers

A method for n-wise registration and mosaicing of partial prints

Automatic identification of fingerprint inpainting target areas

The Logistic Random Field — A convenient graphical model for learning parameters for MRF-based labeling

Secure digital communications via biometric key generation

Fingerprint matcher using iterative process and related methods