Abstract: Systems and methods for setting the white balance of an image are described. Embodiments of the systems and methods may receive image data comprising a plurality of exposures, generate a plurality of white balance values based on merge information from a high dynamic range (HDR) merge of the exposures, and adjust a white balance of each pixel of the image data based on the white balance values.

Abstract: Systems and methods for setting the white balance of an image are described. Embodiments of the systems and methods may receive image data comprising a plurality of exposures, generate a plurality of white balance values based on merge information from a high dynamic range (HDR) merge of the exposures, and adjust a white balance of each pixel of the image data based on the white balance values.

Abstract: A method of rectifying stereo images includes providing a plurality of pairs of sets of keypoints extracted from a pair of current stereo images and from a pair of previous stereo images wherein each pair of stereo images includes a left image and a right image respectively obtained from a left camera and a right camera; providing a plurality of pairs of sets of next-choice-match points extracted from the pair of current stereo images and the pair of previous stereo images; finding one or more anchor points in a left previous image; finding a right linking point which is the corresponding keypoint in the right previous image, and a left linking point which is the corresponding keypoint in the left current image; finding a closing point; and calculating a cost from the right linking point, the left linking point, and the closing point.

Abstract: A method of rectifying stereo images includes providing a plurality of pairs of sets of keypoints extracted from a pair of current stereo images and from a pair of previous stereo images wherein each pair of stereo images includes a left image and a right image respectively obtained from a left camera and a right camera; providing a plurality of pairs of sets of next-choice-match points extracted from the pair of current stereo images and the pair of previous stereo images; finding one or more anchor points in a left previous image; finding a right linking point which is the corresponding keypoint in the right previous image, and a left linking point which is the corresponding keypoint in the left current image; finding a closing point; and calculating a cost from the right linking point, the left linking point, and the closing point.

Abstract: A method of evaluating motion estimation between a pair of digitized images includes receiving a distance map between a source block in a source image and all the blocks in a search area in a target image, scanning each column of the distance map, and saving indices of a minimum distance value for each column, scanning each row of the distance map, and saving indices of a minimum distance value for each row, locating candidate lines that pass through at least some local minima points that correspond to locations in the distance map of the minimum distance value in each of the columns or the minimum distance value in each of the rows determining a confidence level for each candidate line that passes through at least some of the local minima points, and selecting those candidate lines whose confidence level is greater than a predetermined threshold value.

Abstract: A method of matching stereo images includes: determining, first information indicating which orientations are dominant in each pixel location of a right image and a left image among the stereo images; determining second information indicating for each pixel location in the right image and the left image whether it lies on a pure horizontal edge; detecting points in the left image and the right image based on the first information and the second information; and performing a matching on the left image and the right image using the detected points to estimate a sparse disparity map.

Abstract: Exemplary embodiments of the invention as described herein generally provide for detecting the displacement of feature(s) within a visual image in cases where pattern matching fails due to the existence of aperture(s) caused for example by external condition(s) encountered in recording such an image over time. Technique(s) are disclosed for detecting the difference between displacement of a geometric feature of an object appearing within an image (e.g., an edge or smooth surface) that has an aperture and another feature (e.g., a corner) that does not since it is not symmetrically invariant.

Abstract: A method of matching stereo images includes: determining, first information indicating which orientations are dominant in each pixel location of a right image and a left image among the stereo images; determining second information indicating for each pixel location in the right image and the left image whether it lies on a pure horizontal edge; detecting points in the left image and the right image based on the first information and the second information; and performing a matching on the left image and the right image using the detected points to estimate a sparse disparity map.

Abstract: A method of evaluating motion estimation between a pair of digitized images includes receiving a distance map between a source block in a source image and all the blocks in a search area in a target image, scanning each column of the distance map, and saving indices of a minimum distance value for each column, scanning each row of the distance map, and saving indices of a minimum distance value for each row, locating candidate lines that pass through at least some local minima points that correspond to locations in the distance map of the minimum distance value in each of the columns or the minimum distance value in each of the rows determining a confidence level for each candidate line that passes through at least some of the local minima points, and selecting those candidate lines whose confidence level is greater than a predetermined threshold value.

Abstract: Exemplary embodiments of the invention as described herein generally provide for detecting the displacement of feature(s) within a visual image in cases where pattern matching fails due to the existence of aperture(s) caused for example by external condition(s) encountered in recording such an image over time. Technique(s) are disclosed for detecting the difference between displacement of a geometric feature of an object appearing within an image (e.g., an edge or smooth surface) that has an aperture and another feature (e.g., a corner) that does not since it is not symmetrically invariant.

Abstract: An example embodiment discloses a method of determining an estimated depth of an object in an image. The method includes determining an estimated shear due to a rolling shutter image sensor in a plurality of regions of a current frame of the image and determining the estimated depth based on the estimated shear.

Abstract: An example embodiment discloses a method of determining an estimated depth of an object in an image. The method includes determining an estimated shear due to a rolling shutter image sensor in a plurality of regions of a current frame of the image and determining the estimated depth based on the estimated shear.

Abstract: A motion estimation method includes grouping row blocks of a first frame into a first plurality of banks and grouping row blocks of a second frame into a corresponding second plurality of banks, respectively; calculating the normalized cross correlation (NCC) or the sum of absolute difference (SAD) between the banks of the first frame and the banks of the second frame; detecting the local maxima from the NCC or the local minima from the SAD; estimating a first relative displacement between corresponding reference row blocks of the first frame and of the second frame; calculating motion coefficients using first relative displacements; and estimating second relative displacements using the motion coefficients.

Abstract: A method of detecting an ROI is provided. The method includes calculating energy of each of unit blocks constituting an image frame, detecting at least one interest block having energy higher than a threshold value among the unit blocks, forming initial ROIs by dividing the image frame, and removing a medium ROI among the initial ROIs.

Abstract: A motion vector processing device includes a motion vector detection unit, a transforming unit, and a clustering processing unit. The motion vector detection unit detects motion vectors from a current image frame and a reference image frame. The transforming unit transforms the detected motion vectors into corresponding points in a detection space. The clustering processing unit clusters the corresponding points in at least one cluster each having a bandwidth. The clustering processing unit analyzes a relation between a bandwidth of each of at least one cluster and a total number of clusters of the at least one cluster.

Abstract: A motion estimation method includes grouping row blocks of a first frame into a first plurality of banks and grouping row blocks of a second frame into a corresponding second plurality of banks, respectively; calculating the normalized cross correlation (NCC) or the sum of absolute difference (SAD) between the banks of the first frame and the banks of the second frame; detecting the local maxima from the NCC or the local minima from the SAD; estimating a first relative displacement between corresponding reference row blocks of the first frame and of the second frame; calculating motion coefficients using first relative displacements; and estimating second relative displacements using the motion coefficients.

Abstract: In a method for rectifying rolling shutter effects, a horizontal translation and a vertical translation of a second image relative to a first image are estimated, and the estimated horizontal and vertical translations are filtered using a Kalman filter. Rolling shutter distortion of the second image is rectified according to the filtered horizontal and vertical translations.