Arcsoft Corporation Limited

Jiangen Cao

Jin Wang

Shu Li

Yingfeng Su

Congxiu Wang

Dongren Chen

Guogang Wang

Yanying Sun

Huaqi Zhang

Hanguang Dan

Sheng Han

A method for concatenating MPEG-4 or H.263 segments is provided. The method includes extracting and comparing the VOL or picture header information of each segment. If the two segments have the same VOL or picture header information, the method merges the two segments directly by appending one segment to another to form a new bit stream. If the two segments have the different resolution, the method completely decodes and re-encodes the segments and then merges the segments. Otherwise the method takes the VOL or picture header information of one segment as reference to produce the new bit stream. The method parses the other segment frame at a time, re-encapsulates it according to the reference VOL or picture header information, and appends it to the one segment. During the process, the method only uses VLD to parse the other segment without using the tools of DCT, IDCT, ME, and MC.

Video concatenation for MPEG-4 and H.263 bit streams

A method for detecting a facial area on a color image includes (a) placing a search window on the color image, (b) determining if a center pixel of the search window is a skin color pixel, indicating that the search window is a possible facial area candidate, (c) applying a 3-rectangle filter to the search window to determine if the search window is a possible facial area candidate, (d) applying a 4-rectangle filter to the search window to determine if the search window is a possible facial area candidate, (e) if steps (b), (c), (d) all determine that the search window is a possible facial area candidate, applying an AdaBoost filter to the search window to determine if the search window is a facial area candidate, and (f) if step (e) determines that the search window is a facial area candidate, saving the location of the search window.

Face detection on mobile devices

A method is provided for approximating a quadratic Bezier curve represented by two anchor points and a control point that are fixed-point binary numbers. If a first flatness of a line formed between the two anchor points of the Bezier curve is less than a threshold, the method includes replacing the Bezier curve with an edge between the two anchor points. If the first flatness is not less than the threshold, the method includes subdividing the Bezier curve into halves and determining a second flatness of one of the halves. The first flatness is approximated by a triangular area formed by the two anchor points and the control point, and the triangular area is approximated by multiplication and subtraction operations. The second flatness of the half is approximated as the triangular area divided by 8, where the division by 8 is determined by shifting 3 bits to the right.

Graphic engine for approximating a quadratic bezier curve in a resource-constrained device

Locating an eye includes generating an intensity response map by applying a 3-rectangle filter and applying K-mean clustering to the map to determine the eye. Locating an eye corner includes applying logarithm transform and grayscale stretching to generate a grayscale eye patch, generating a binary map of the patch by using a threshold based on a histogram of the patch, and estimating the eye corner by averaging coordinates weighted by minimal eigenvalues of spatial gradient matrices in a search region based on the binary map. Locating a mouth corner includes generating another intensity response map and generating another binary map using another threshold based on another histogram of the intensity response map. Locating a chin or a cheek includes applying angle constrained gradient analysis to reject locations that cannot be the chin or cheek. Locating a cheek further includes removing falsely detected cheeks by parabola fitting curves through the cheeks.

Facial feature detection on mobile devices

A data structure, termed a pyramid, representing levels of data is disclosed. A method to produce the data structure is also disclosed. The pyramid data structure includes a series of levels, termed floors. In turn, the each floor is comprised of numerous cells. The data structure may be cached or can be stored in a file or database for run-time access by distributed applications.

Process and data structure for providing required resolution of data transmitted through a communications link of given bandwidth

A method for rasterizing an edge includes determining an area projected by the edge in a pixel it touches. The projected area is first set equal to a unit area projected by the edge in one unit step. If a starting point of the edge is located in the pixel, an excessive projected area attributed to the edge before the starting point is subtracted from the projected area. If an end point of the edge is located in the pixel, another excessive projected area attributed to the edge after the end point is subtracted from the unit area from the projected area. If the edge crosses the pixel, then a carryover projected area attributed to the edge in a preceding pixel is added to the projected area.

Growing community of inventors

Dublin, CA, United States of America

Jiangen Cao