Technion Research & Development Foundation Limited

Lucent Technologies Inc.

Alfred M Bruckstein

Ron Kimmel

Thomas J Richardson

Arun N Netravali

Robert J Holt

Israel A Wagner

Michal Aharon

Ilya Blayvas

Tal Nir

Aaron Wetzler

Thomas S Huang

Michael Elad

Vladimi Yanovski

Roberto Mecca

A method of processing a set of images of a scene is disclosed. The set is represented by a plurality of picture-elements defining an image plane, and the images are characterized by different light intensity spatial distributions but the same viewpoint relative to the scene. The method comprises: for each of at least some picture-elements, independently selecting a direction parallel to the image plane, constructing a weighted vector field based on the selected direction, and calculating a height of a surface in the scene perpendicular to the image plane based on the constructed weighted vector field. The method further comprises reconstructing a three-dimensional spatial representation of the surface based on the heights; and transmitting the spatial representation to a computer readable medium.

3D reconstruction from photometric stereo with shadows

An optical flow estimation process based on a spatio-temporal model with varying coefficients multiplying a set of basis functions at each pixel. The benefit of over-parameterization becomes evident in the smoothness term, which instead of directly penalizing for changes in the optic flow, accumulates a cost of deviating from the assumed optic flow model. The optical flow field is represented by a general space-time model comprising a selected set of basis functions. The optical flow parameters are computed at each pixel in terms of coefficients of the basis functions. The model is thus highly over-parameterized, and regularization is applied at the level of the coefficients, rather than the model itself. As a result, the actual optical flow in the group of images is represented more accurately than in methods that are known in the art.

Over-parameterized variational optical flow method

A signal processing system adapted for sparse representation of signals is provided, comprising: (i) one or more training signals; (ii) a dictionary containing signal-atoms; (iii) a representation of each training signal using a linear combination of said dictionary's signal-atoms; (iv) means for updating the representation of the training signal; (v) means for updating the dictionary one group of atoms at a time, wherein each atom update may include all representations referring to said updated atom; and (vi) means for iterating (iv) and (v) until a stopping rule is fulfilled. The system uses the K-SVD algorithm for designing dictionaries for sparse representation of signals.

System and method for designing of dictionaries for sparse representation

A method of coloring a plurality of cells using a predetermined fill-color, the cells occupying a region in a grid of cells thereby defining a plurality of region-cells, the method comprising: (a) dividing the grid into a plurality of elementary-regions, each comprising at least one cell; (b) visiting one elementary-region at a time and determining, for each cell of the elementary-region, whether the cell is a region-cell; (c) writing color data in at least one region-cell of each elementary-region using a color selected from a color-space so as to store information in the elementary-region, the information directing a backtracking path in the region, thereby providing a mapped elementary-region; and (d) using the information for visiting all the mapped elementary-regions along the backtracking path and for writing color data using the fill-color in each region-cell of each of the mapped elementary-region along the backtracking path; thereby coloring all the region-cells by the fill-color.

Memory free method of region filling

Data processing apparatus is provided for evaluating answers to respective query items considered to be represented by respective points within a region of feature space, and the apparatus comprises an input () which receives such a query item. The region is subdivided into subregions according to at least first and second subdivisions. A subregion identifying portion () identifies, for each such subdivision of the region, which subregion of the subdivision contains the point representing the received query item. A partial answer retrieval portion () has access when the apparatus is in use to a store () of precalculated partial answers for at least some the subregions of the subdivisions, and retrieves from the store the partial answers for the or each identified subregion that is present in the store. An answer calculation portion () calculates an answer to the received query item based on the retrieved partial answers, and an output () outputs the calculated answer.

Machine learning by construction of a decision function

A method and system for watermarking and recovering a watermarked digitized image for subsequent authentication. The digitized image comprises spatial domain data signals reproduced from an original image. A transformer transforms the spatial domain data signals into frequency domain data signals, including respective magnitude and phase data signal components. Information is embedded into the magnitude data signals to develop modified data signals and the modified data signals are converted from the frequency domain to the spatial domain to generate the watermarked digitized image. For recovery, a second transformer transforms the spatial domain data signals comprising the respective original digitized image and watermarked digitized image into respective frequency domain dataset signals. A comparator responsive to the two transformer outputs compares the respective frequency domain dataset signals to identify the watermark and verify the authenticity of the image reproduction.

Transform domain image watermarking method and system

In a communications system, an image to be transmitted is holographically represented in accordance with the invention. Relying on a portion of the holographic representation selected from any part thereof, a receiver can provide an overall version of the transmitted image. This overall version is successively refined as the size of the aggregate portion of the representation received by the receiver increases.

Technique for holographic representation of images

Pose estimation of a camera is facilitated by including in the scene a space fiducial, an object that has the property that its appearance as viewed by the camera uniquely determines the direction from which it is viewed.

Process for pose estimation of a camera viewing an image scene

An omnidirectional or panoramic viewer/projector uses a single camera and a mirror with a curved surface. The curved mirror provides a single virtual optical center or viewpoint.

Growing community of inventors

Haifa, Israel

Alfred M Bruckstein