Ramakrishna Kakarala

US Patent:

6819801, Nov 16, 2004

Filed:

Jun 19, 2001

Appl. No.:

09/885389

Inventors:

Ramakrishna Kakarala - Sunnyvale CA
Izhak Baharav - San Jose CA

Assignee:

Agilent Technologies, Inc. - Palo Alto CA

International Classification:

G06K 936

US Classification:

382240, 382166, 382269

Abstract:

A system and method is provided for processing a demosaiced image using a color aliasing artifact reduction (CAAR) algorithm in order to reduce color aliasing artifacts. The CAAR algorithm computes the L level wavelet transform for the demosaiced color planes R, G and B. Thereafter, the CAAR algorithm estimates the correct color value at each pixel location for the colors not associated with that pixel location. For example, to determine the green value at red pixel locations, the CAAR algorithm performs an inverse wavelet transform using the green approximation signal and the red detail signals. This process is repeated for each of the colors (e. g. , green values at blue pixel locations, red values at green pixel locations, etc. ). In addition, the CAAR algorithm performs an inverse wavelet transform on each of the color planes themselves, so that the pixel values of the color associated with each pixel location are not altered. Thereafter, the inverse wavelet transform of each color plane is combined with the inverse wavelet transform of each of the estimated color values for that color plane to produce correlated R, G and B color planes.

US Patent:

6989862, Jan 24, 2006

Filed:

Aug 23, 2001

Appl. No.:

09/938438

Inventors:

Izhak Baharav - Palo Alto CA, US
Ramakrishna Kakarala - Sunnyvale CA, US
Dietrich W. Vook - Menlo Park CA, US

Assignee:

Agilent Technologies, Inc. - Palo Alto CA

International Classification:

H04N 5/335

US Classification:

348273

Abstract:

A system and method for processing mosaiced or raw data images operates to concurrently demosaic and resize the mosaiced images in a combined process. The combined demosaic/resize process allows the system to perform demosaicing and resizing more efficiently than conventional systems, which perform these processes separately and sequentially. Furthermore, the combined demosaic/resize process allows the system to produce demosaiced and resized images of higher quality as compared to demosaiced and resized images produced by the conventional systems.

US Patent:

7015961, Mar 21, 2006

Filed:

Aug 16, 2002

Appl. No.:

10/222150

Inventors:

Ramakrishna Kakarala - Sunnyvale CA, US

International Classification:

H04N 9/64
H04N 5/335
H04N 5/228
G06K 9/32

US Classification:

348246, 348273, 3482221, 382300

Abstract:

A digital image system and method for combining bad pixel correction and demosaicing in a single process is provided by interpolating sensor values for pixels immediately spatially adjacent to the current pixel being examined to detect defective pixels, and using the interpolated values for demosaicing. If the sensor value of the current pixel being examined is outside of a range of sensor values determined from the interpolated values by more than a configurable threshold amount, the current pixel is considered defective, and replaced using an estimated value from the neighboring pixels. The interpolated values calculated for use in detecting bad pixels can further be used as the interpolated values for demosaicing purposes.

US Patent:

7019826, Mar 28, 2006

Filed:

Mar 20, 2003

Appl. No.:

10/392990

Inventors:

Dietrich W. Vook - Los Altos CA, US
Izhak Baharav - San Jose CA, US
Xuemei Zhang - Mountain View CA, US
Ramakrishna Kakarala - Santa Clara CA, US
Richard L. Baer - Los Altos CA, US

Assignee:

Agilent Technologies, Inc. - Palo Alto CA

International Classification:

G01N 21/00
G01N 21/88
G06K 9/00

US Classification:

3562371, 3562374, 25055934, 382150

Abstract:

A three-dimensional optical inspection system reconstructs a three-dimensional image of the shape of the surface of an at least partially specular object resident on a printed circuit board by capturing two or more two-dimensional images of the object under different illumination configurations. The diffuse reflection, as well as the specular reflection can be used to reconstruct the three-dimensional image using any reconstruction method, such as photometric stereo. The different illumination configurations can be achieved using an illumination source including light-emitting elements arranged in concentric circular arrays, in which each of the circular arrays is divided into sections. Each section is independently controlled to selectively activate the sections to illuminate the object in a pre-established illumination pattern.

US Patent:

7079705, Jul 18, 2006

Filed:

Oct 30, 2002

Appl. No.:

10/284696

Inventors:

Xuemei Zhang - Mountain View CA, US
Ramakrishna Kakarala - Sunnyvale CA, US
Todd S. Sachs - Palo Alto CA, US

Assignee:

Agilent Technologies, Inc. - Palo Alto CA

International Classification:

G06K 9/00
G06K 9/32
G06K 9/36
H04N 1/46

US Classification:

382280, 382162, 382300, 358525

Abstract:

An image processing system and demosaicing method are provided to calculate estimated missing color sensor values in an image using a linear prediction from the raw color sensor value at the current pixel location. The raw image is divided into regions of sensor values, and the linear relations between color planes for each region are determined by a regression method that calculates the degree to which different color planes co-vary within each region The missing color sensor values per region are calculated as a scaled and shifted version of the raw color sensor values using linear regression coefficients determined from the local linear regression process.

US Patent:

7088392, Aug 8, 2006

Filed:

Aug 27, 2001

Appl. No.:

09/940825

Inventors:

Ramakrishna Kakarala - Sunnyvale CA, US
Izhak Baharav - San Jose CA, US

International Classification:

H04N 5/335

US Classification:

348272, 382167, 348252, 3482221

Abstract:

An adaptive demosaicing method interpolates images based on color edge detection and neighborhood voting. The adaptive demosaicing algorithm uses a voting scheme to determine the direction of interpolation at each missing luminance pixel location. Each color plane votes either horizontal or vertical based on a comparison between the horizontal and vertical components of the degree of change (i. e. , gradient, Laplacian or other measure of the degree of change) in that color plane. Votes are counted from the neighborhood pixels as well as from measurements taken at the pixel location itself. Once the luminance plane is fully interpolated, the chrominance planes are filled in by simple bilinear or median interpolation of difference chrominance values. Enhancements to the adaptive demosaicing algorithm permit adaptive smoothing and sharpening.

US Patent:

7092578, Aug 15, 2006

Filed:

Oct 23, 2001

Appl. No.:

10/001508

Inventors:

Ramakrishna Kakarala - Sunnyvale CA, US
Jonathan S. Gibson - Aurora OR, US

Assignee:

Agilent Technologies, Inc. - Palo Alto CA

International Classification:

G06K 9/36
H04N 1/32

US Classification:

382251, 382239, 382246, 35824002

Abstract:

Previously-unused slots in a Huffman code table associated with a Joint Photographic Experts Group (JPEG) image file are associated with various quantization matrices (Q matrices) that are used to quantize data blocks of the JPEG image file. Huffman codes associated with the various Q matrices permit the particular Q matrix used to quantize a given data block to be signaled by a decoder as an end-of-block (EOB) code. The EOB codes and the Huffman code table are sent with the JPEG image file. Upon decoding of the image file, a standard JPEG decoder reads each of the EOB codes as a standard JPEG EOB code and does not vary the Q matrix. A modified decoder reads from each of the EOB codes which Q matrix was used to encode each particular data block of the image and uses that Q matrix to dequantize the data block.

US Patent:

7155066, Dec 26, 2006

Filed:

May 31, 2001

Appl. No.:

09/871408

Inventors:

Izhak Baharav - San Jose CA, US
Ramakrishna Kakarala - Sunnyvale CA, US

Assignee:

Agilent Technologies, Inc. - Santa Clara CA

International Classification:

G06K 9/46

US Classification:

382248, 382275

Abstract:

A system and method for processing mosaiced images utilizes a compression-aware demosaicing process that takes into consideration a subsequent compression process. The compression-aware demosaicing process is performed using a compression-considered demosaicing operator that incorporates a color space conversion operator and a frequency-based transformation operator, which are typically associated with the compression process. Consequently, the overall efficiency of the system and method is significantly increased. Furthermore, the compression-aware demosaicing process produces artifacts that complement the artifacts produced by the subsequent compression process such that the artifacts are less visible in the final color images, which increases the quality of the final color images.