Dan C Lelescu

7308131, Dec 11, 2007

Dec 3, 2002

10/308828

Dan Lelescu - Morgan Hill CA, US
Frank Jan Bossen - San Jose CA, US

NTT DoCoMo, Inc. - Tokyo

G06K 9/00
G06K 9/32
H04N 7/00

382154, 382293, 348 36

A method for representing an omni directional-like image by capturing an omnidirectional-like image and determining a representation format that is commonly applicable to any image acquisition system that is able to capture the omnidirectional-like image. The captured omnidirectional-like image is mapped onto the determined representation format.

7924918, Apr 12, 2011

Nov 28, 2005

11/288822

Dan Lelescu - Morgan Hill CA, US
Frank Bossen - Plan-les Ouates, CH

NTT Docomo, Inc. - Tokyo

H04N 11/02
H04N 11/04

37524012, 37524024, 37524016, 37524025

A method, apparatus and article of manufacture for performing temporal prediction are described. In one embodiment, a decoding method includes, but is not limited to, generating a prediction for a sample in one of multiple blocks using transformation parameters associated with the one block and transformation parameters associated with at least one other block in the group of blocks and adding residual data to the prediction to obtain a decoded sample.

8050331, Nov 1, 2011

May 18, 2006

11/437515

Dan Lelescu - Morgan Hill CA, US
Frank Bossen - Plan-les Ouates, CH

NTT DoCoMo, Inc. - Tokyo

H04N 7/12
H04N 11/02

37524029, 37524018

A method and apparatus is disclosed herein for encoding and/or decoding video frame data. In one embodiment, the video coder comprises a noise filtering module to operate on transformed frame data and perform signal estimation on a plurality of transform coefficients by estimating signal power for each of the plurality of transform coefficients, comparing the signal power of said each coefficient with at least one threshold, and setting the value of said each transform coefficient based, at least in part, on results of comparing the signal power to the at least one threshold.

8131097, Mar 6, 2012

May 28, 2008

12/128234

Dan Lelescu - Morgan Hill CA, US
Robert Mullis - Soquel CA, US
Pravin Rao - San Jose CA, US
Kartik Venkataraman - San Jose CA, US
Cheng Lu - Santa Clara CA, US
Junqing Chen - San Jose CA, US

Aptina Imaging Corporation - George Town

G06K 9/40

382254

A method and apparatus are disclosed for restoring an image captured through an extended depth-of-field lens. Preprocessed data relating to image degradation is stored and used during an image restoration process.

8135233, Mar 13, 2012

May 22, 2008

12/153684

Dan Lelescu - Morgan Hill CA, US

Aptina Imaging Corporation - George Town

G06K 9/40

382275

An imager and method for operating the imager. The imager comprises a pixel array for producing an image signal and an image processor configured to restore images degraded by inter-channel degradations and intra-channel degradations in all channels of the image simultaneously by diagonalizing an image degradation matrix and componentizing the diagonalized image degradation matrix and multi-channel image.

8526754, Sep 3, 2013

May 28, 2009

12/474215

Brian Keelan - Boulder Creek CA, US
Dan Lelescu - Morgan Hill CA, US

Aptina Imaging Corporation - George Town

G06K 9/40

382255, 250201, 250231, 324307, 324309, 345419, 345413, 3482221, 348335, 348345, 348348, 348352, 382108, 382129, 382131, 382254, 382260, 382261, 382284

An electronic device may have a digital image sensor for capturing an image and may have storage and processing circuitry for processing the image to produce a corresponding enhanced depth-of-field image. Multiple image areas, within a given image may be processed. As each image area is processed, a signal-to-noise ratio may be computed. A kernel parameter characteristic curve may be used to select values to be used for kernel parameters for each image area based on the computed signal-to-noise ratio for that image area. Kernels may be generated for each image area and the generated kernels may be convolved with their respective image areas to modify the image areas. This allows each image area to be selectively sharpened or smoothed based on its signal-to-noise ratio. Image areas with low and high signal-to-noise ratios may be sharpened less those with moderate signal-to-noise ratios to enhance depth of field in the image.

8587703, Nov 19, 2013

Dec 1, 2009

12/628604

Dan Lelescu - Morgan Hill CA, US
Chihsin Wu - San Jose CA, US
David R. Pope - Fremont CA, US
Cheng Lu - Santa Clara CA, US

Aptina Imaging Corporation - Georgetown, Grand Cayman

H04N 9/64
G06K 9/40
H04N 3/14

348251, 382254, 348272

Systems and methods are provided for focus recovery of multi-channel images. Control circuitry of an imaging system can restore an image by removing image blurring introduced by the lens, sensor noise introduced by the sensor, and a signal offset between multiple channels of the image. In some embodiments, the control circuitry can calculate one or more estimates of a signal offset of multiple observed signals. Using statistics generated from offset-removed signals, the control circuitry can generate one or more recovery kernels which can be applied to offset-removed signals to generate recovered signals. In other embodiments, instead of explicitly removing a signal offset from each observed signal, the control circuitry can implicitly remove the signal offset when calculating the first and second order statistics of one or more observed signals.

8619082, Dec 31, 2013

Aug 21, 2013

13/972881

Kartik Venkataraman - San Jose CA, US
Gabriel Molina - Palo Alto CA, US
Dan Lelescu - Morgan Hill CA, US

Pelican Imaging Corporation - Mountain View CA

G06T 15/10
G06K 9/00

345427, 382154

Systems in accordance with embodiments of the invention can perform parallax detection and correction in images captured using array cameras. Due to the different viewpoints of the cameras, parallax results in variations in the position of objects within the captured images of the scene. Methods in accordance with embodiments of the invention provide an accurate account of the pixel disparity due to parallax between the different cameras in the array, so that appropriate scene-dependent geometric shifts can be applied to the pixels of the captured images when performing super-resolution processing. In several embodiments, detecting parallax involves using competing subsets of images to estimate the depth of a pixel location in an image from a reference viewpoint. In a number of embodiments, generating depth estimates considers the similarity of pixels in multiple spectral channels. In certain embodiments, generating depth estimates involves generating a confidence map indicating the reliability of depth estimates.