Xerox Corporation

The Penn State Research Foundation

USA As Represented by Secretary of the Navy

Conduent Business Services, LLC

Vishal Monga

Raja Bala

Zhigang Fan

Shen-Ge Wang

Claude Fillion

Gaurav Sharma

Xuan Mo

Aaron Michael Burry

Orhan Bulan

Ramesh Nagarajan

Reiner Eschbach

Wencheng Wu

Jon S McElvain

Michael David Shepherd

Dale Ellen Gaucas

Michael Branciforte

Rajinderjeet Singh Minhas

Judith E Stinehour

Jose Antonio Rodriguez-Serrano

Raghu G Raj

Manu Parmar

Divyanshu Vats

John Mckay

Claude S Fillion

Systems and methods are provided for imaging that demonstrably outperform previous approaches (especially compressive sensing based approaches). Embodiments of the present disclosure provide and solve an imaging cost function via a stochastic approximation approach. By doing so, embodiments of the preset disclosure provide a significant means of generalization and flexibility to adapt to different application domains while being competitive in terms of computational complexity.

Tomographic imaging using fast stochastic hierarchical Bayesian systems and methods

A method and system for detecting anomalies in video footage. A training dictionary can be configured to include a number of event classes, wherein events among the event classes can be defined with respect to n-dimensional feature vectors. One or more nonlinear kernel function can be defined, which transform the n-dimensional feature vectors into a higher dimensional feature space. One or more test events can then be received within an input video sequence of the video footage. Thereafter, a determination can be made if the test event(s) is anomalous by applying a sparse reconstruction with respect to the training dictionary in the higher dimensional feature space induced by the nonlinear kernel function.

Anomaly detection using a kernel-based sparse reconstruction model

Methods, systems, and processor-readable media for video anomaly detection based upon a sparsity model. A video input can be received and two or more diverse descriptors of an event can be computed from the video input. The descriptors can be combined to form an event matrix. A sparse reconstruction of the event matrix can be performed with respect to an over complete dictionary of training events represented by the diverse descriptors. A step can then be performed to determine if the event is anomalous by computing an outlier rejection measure on the sparse reconstruction.

Video anomaly detection based upon a sparsity model

Methods and systems for detecting anomalies in transportation related video footage. In an offline training phase, receiving video footage of a traffic location can be received. Also, in an offline training phase, event encodings can be extracted from the video footage and collected or compiled into a training dictionary. One or more input video sequences captured at the traffic location or a similar traffic location can be received in an online detection phase. Then, an event encoding corresponding to the input video sequence can be extracted. The event encoding can be reconstructed with a low rank sparsity prior model applied with respect to the training dictionary. The reconstruction error between actual and reconstructed event encodings can then be computed in order to determine if an event thereof is anomalous by comparing the reconstruction error with a threshold.

Detecting multi-object anomalies utilizing a low rank sparsity model

Methods and systems for automatically detecting multi-object anomalies at a traffic intersection utilizing a joint sparse reconstruction model. A first input video sequence at a first traffic location can be received and at least one normal event involving P moving objects (where P is greater than or equal to 1) can be identified in an offline training phase. The normal event in the first input video sequence can be assigned to at least one normal event class and a training dictionary suitable for joint sparse reconstruction can be built in the offline training phase. A second input video sequence captured at a second traffic location similar to the first traffic location can be received and at least one event involving P moving objects can be identified in an online detection phase.

Method and system for automatically detecting multi-object anomalies utilizing joint sparse reconstruction model

Methods, systems, and processor-readable media for pruning a training dictionary for use in detecting anomalous events from surveillance video. Training samples can be received, which correspond to normal events. A dictionary can then be constructed, which includes two or more classes of normal events from the training samples. Sparse codes are then generated for selected training samples with respect to the dictionary derived from the two or more classes of normal events. The size of the dictionary can then be reduced by removing redundant dictionary columns from the dictionary via analysis of the sparse codes. The dictionary is then optimized to yield a low reconstruction error and a high-interclass discriminability.

Dictionary design for computationally efficient video anomaly detection via sparse reconstruction techniques

This disclosure provides methods, systems and apparatus for jointly optimizing node locations and corresponding output value of a color look-up-table (LUT) associated with an imaging device. According to one exemplary method, initially a set of LUT node locations are generated by solving a first optimization problem and subsequently, a set of LUT node values are generated by solving a second optimization problem.

Methods, systems and apparatus for jointly optimizing node locations and corresponding output values of a color look-up-table (LUT)

Systems and methods are described that facilitate reducing a number of patches used in characterizing a color halftone printer via a binary color printer model. A binary printer model involves printing of a fundamental set of color binary patterns that encompass all possible halftone outputs. A k-center clustering technique is employed to automatically find and eliminate redundancies in the initial set of binary color patterns. Once the number of patches is reduced to an acceptable number, a multiplicative reflectance model is applied that better approximates the physical process and therefore improves accuracy.

Cost-effective binary printer models for multi-color printers by improved reflectance modeling and patch clustering

A method and system is disclosed for characterizing a color scanner comprising generating a halftone-independent target of color patches, printing the target on a color hardcopy device, measuring the target to obtain device-independent color values, scanning the target to obtain scanner color values, and building a scanner profile that relates scanned color values to device-independent color values.

Halftone-independent scanner profiling

A system/method of color match assessment for electronic documents includes receiving digital data defining a composite electronic document including a raster image object having an edge and a color graphics object bordering the edge of the raster image object. The pixel color values defining the edge of the raster image object are processed to estimate a local color variance of the pixel color values. The local color variance is used to determine if the edge can be color matched to the bordering color graphics object. If the edge can be color matched, a match color for the edge is derived. The match color is associated with the digital data defining the electronic document so that a downstream object color match system can use the match color as needed.

Colorimetric matching the perception of a digital data file to hardcopy legacy

What is disclosed is a resizing method that utilizes segmentation information to classify objects found within a document and then selects the most appropriate resizing technique for each identified object. The present method employs readily available document parsers to reliably extract objects. e.g. text, background, images, graphics, etc., which compose the document. Information obtained from a document parser is utilized to identify the document components for classification. The extracted objects are then classified according to their object type. Each of classified objects are then resized using a resizing technique having been pre-selected for the object type based on their respective abilities to resize certain types of document content over other resizing techniques. The present method advantageously extends smart or content-based scaling and is especially useful for N-up or variable-information printing. The present method finds its intended uses in enhancing N-up and handout options currently provided in a variety of print-drivers.

Object based adaptive document resizing

This disclosure provides methods, apparatus and systems for performing image processing regression for approximating multidimensional color transformation. According to an exemplary method, a shaping matrix is selected to minimize a cost function associated with a local linear regression representation of the color transformation. In addition, an alternating least squares algorithm is utilized to jointly optimize regression and shaping parameters.

Local regression methods and systems for image processing systems

A system resizes content within a document that includes a document segmenter that receives a document that contains content. The document segmenter analyzes the content within the document and segments the content into a plurality of object types. An object priority applicator determines a class value associated with each object type. A location scaler identifies a datum point for each object type within the document, wherein each datum point maintains a relative location to one another regardless of document resizing. An object sizing component resizes each object based at least in part upon the class value.

Systems and methods to resize document content

A method and apparatus are provided for compensating for spatial non-uniformities in a printer by deriving a true spatial non-uniformity tone response curve (TRC) that characterizes the printer in terms of color output variation for each addressable pixel location in a spatial range. The 'true average' tone response curve is determined for a color channel. A prediction of the true response as a function of the spatial location is derived by printing and scanning a specially designed halftone-independent target of binary patterns. The predicted tone response curve for each color channel and halftone is predicted using a binary printer model, wherein the 'predicted tone response curve' provides a model based approximation of the actual response for each addressable pixel location in the spatial range. Also stored is an &#x201c;average predicted tone response&#x201d; by averaging the &#x201c;predicted tone response curve&#x201d; over the spatial range of the printer.

Halftone independent correction of spatial non-uniformities

A system and method for resizing a digitally represented color image are presented. A color image with pixels defined by luminance and at least one chrominance value is received. For each pixel of the color image, a luminance spatial variation and respective chrominance spatial variations in the respective neighborhood of the each pixel are computed. The luminance spatial variation and the respective chrominance spatial variations are combined to produce a respective importance value for each pixel. Selected pixels are identified based upon their respective importance values and are removed by seam carving of the color image. The seam carving identifies seams of pixels based upon the respective importance values of pixels within the seams of pixels to create a resized color image. The resized color image is produced to an image output device.

Growing community of inventors

Webster, NY, United States of America

Vishal Monga