Abstract: An image blending apparatus and method thereof are provided. The image blending apparatus includes an image providing module providing a first image with a first overlap region and a second image with a second overlap region, and an image blending module generating a first gradient image of the first image and a second gradient image of the second image, calculating first distance weights of first pixels in the first overlap region of the first gradient image and second distance weights of second pixels in the second overlap region of the second gradient image, blending the first gradient image and the second gradient image into a blended gradient image according to the first distance weights of the first pixels and the second distance weights of the second pixels at respective corresponding positions, and restoring a blended image from the blended gradient image.

Abstract: A method for feature description is provided. The method includes steps of: receiving high dimension data corresponding to a feature patch obtained by a feature extraction algorithm; selecting a plurality of dimension data sets from the high dimension data; comparing different dimension data in each dimension data set to generate a corresponding comparing result for each dimension data set; and generating a binary string according to the comparing results to descript the feature patch.

Abstract: A depth information-based modeling method, a graphic processing apparatus and a storage medium are provided. The depth information-based modeling method includes receiving a plurality of depth image data, each of which has depth information, and a plurality of color images, each of which color information. In the method, a plurality of 3D grids is obtained according to the color information, depth information and a plurality of uniform sampling grids. Each of the uniform sampling grids is further divided into a plurality of sub-grids. At least one point on each of the 3D grids is determined by the depth information, and the triangle meshes are generated according to the points.

Abstract: A depth map aligning method and a system are provided. The depth map aligning method includes the following steps. A view warping is performed to a left-view image and a right-view image of a real-view image according to a corresponding depth map to obtain a right-virtual view image, a left hole information, and a right hole information thereof. A residual difference of pixel, which is not a hole, is obtained by pixel to pixel. If the residual difference of the pixel is larger than a first threshold value, a reverse warping is performed to obtain a coordinate of the pixel. The depth value of the pixel is aligned according to the depth value of one or more pixels in a vicinity of the obtained coordinate.

Abstract: A depth information-based modeling method, a graphic processing apparatus and a storage medium are provided. The depth information-based modeling method includes receiving a plurality of depth image data, each of which has depth information, and a plurality of color images, each of which color information. In the method, a plurality of 3D grids is obtained according to the color information, depth information and a plurality of uniform sampling grids. Each of the uniform sampling grids is further divided into a plurality of sub-grids. At least one point on each of the 3D grids is determined by the depth information, and the triangle meshes are generated according to the points.

Abstract: A depth map aligning method and a system are provided. The depth map aligning method includes the following steps. A view warping is performed to a left-view image and a right-view image of a real-view image according to a corresponding depth map to obtain a right-virtual view image, a left hole information, and a right hole information thereof. A residual difference of pixel, which is not a hole, is obtained by pixel to pixel. If the residual difference of the pixel is larger than a first threshold value, a reverse warping is performed to obtain a coordinate of the pixel. The depth value of the pixel is aligned according to the depth value of one or more pixels in a vicinity of the obtained coordinate.

Abstract: A video search method and an apparatus thereof are provided. In the video search method, the bit streams of a query video file and a plurality of video files to be searched are parsed to obtain a plurality of corresponding motion vectors (MVs). A plurality of corresponding MV maps is constructed in a time domain according to the MVs. Correlations are obtained according to the MV map corresponding to the query video file and the MV maps corresponding to the video files, and a video search result is obtained according to the correlations.

Abstract: A method and system for encoding three-dimensional (3D) video are provided. The method includes: obtaining a depth map of the 3D video, wherein the depth map includes multiple pixels and each of the pixels has a depth value; identifying a first contour of an object in the depth map; changing the depth values according to whether the pixels are located on the first contour to generate a contour bit map; compressing the contour bit map to generate a first bit stream, and decompressing the first bit stream to generate a reconstructed contour bit map; obtaining multiple sampling pixels of the pixels in the object according to a second contour corresponding to the object in the reconstructed contour bit map; and, encoding locations and the depth values of the sampling pixels. Therefore, a compression ratio of the 3D video is increased.

Abstract: A video search method including following steps is provided. Meta-data of a query clip is received, wherein the meta-data includes an index tag and a semantic pattern. One or more candidate clips are retrieved from at least one video database according to the index tag. The semantic pattern is compared with a semantic pattern of each of the candidate clips, and each of the candidate clips is marked as a returnable video clip or a non-returnable video clip according to a comparison result. The candidate clips marked as the returnable video clip are served as a query result matching the query clip. A video search system and a method for establishing a video database are also provided.

Abstract: A compression method and apparatus of depth map in a 3D video are provided. The compression apparatus includes an edge detection module, a homogenizing module, and a compression encoding module. An edge detection is performed on a depth map of a frame in the 3D video. When at least one macroblock in the frame with no object edge passing through is found, a homogenizing processing is performed on the at least one macroblock. And then the depth map is encoded. Therefore, data quantity might be decreased when the depth map is compressed and encoded according to the present disclosure.

Abstract: A method for storing interpolation data is provided. The method uses a buffer in a cache memory and the concept of memory overlap record for storing previously calculated interpolation data, so as to avoid repeated interpolation, thereby decreasing the amount of system operation and the frequency of reading integer points for calculating interpolation from an external memory. Furthermore, a method of data storage for the buffer is provided. The storage method uses the concept of memory address rotation to store interpolation data beyond the boundary of the buffer. Moreover, another storage method is provided, which distributes interpolation data into a plurality of regions in the buffer according to different combinations of decimal coordinates of the interpolation points for economizing the use of memory space and simplifying the search of interpolation data in the buffer.

Abstract: Methods for node distribution in a P2P network system and include the steps of: mapping nodes to a numerical space respectively, wherein each node has a neighborhood table recorded a certain number of nodes and the certain number of nodes are close to the node along one of two sides in increasing and decreasing direction; performing a search process to find a plurality of reference nodes, wherein any two consecutive reference nodes are a first reference node and a second reference node, and the second reference node is the node in the neighborhood table of the first reference node and a distance between the second and the first reference node conforms to a predetermined condition; finding two positioning reference nodes; mapping a target node to a target value, and locating the target node in a position corresponding to the target value and informing the reference nodes of the position.

Abstract: The present invention provides a system for allocating a bitstream of scalable video coding (SVC). The system includes an adapter and an extractor. The adapter is employed for receiving a bitstream produced by a video encoder and producing an extracting parameter of the bitstream according to the feature of a bitstream parameter in the bitstream content. The extractor is employed for extracting and allocating the bitstream according to the above-mentioned extracting parameter. In this way, the present invention is able to achieve the maximum usage efficiency of bandwidth and the optimum frame quality.

Abstract: A video search method including following steps is provided. Meta-data of a query clip is received, wherein the meta-data includes an index tag and a semantic pattern. One or more candidate clips are retrieved from at least one video database according to the index tag. The semantic pattern is compared with a semantic pattern of each of the candidate clips, and each of the candidate clips is marked as a returnable video clip or a non-returnable video clip according to a comparison result. The candidate clips marked as the returnable video clip are served as a query result matching the query clip. A video search system and a method for establishing a video database are also provided.

Abstract: A video search method and an apparatus thereof are provided. In the video search method, the bit streams of a query video file and a plurality of video files to be searched are parsed to obtain a plurality of corresponding motion vectors (MVs). A plurality of corresponding MV maps is constructed in a time domain according to the MVs. Correlations are obtained according to the MV map corresponding to the query video file and the MV maps corresponding to the video files, and a video search result is obtained according to the correlations.

Abstract: A method for packing a bitstream of scalable video coding (SVC) is provided. The method includes the following steps. First, a SVC bitstream is obtained, wherein the SVC bitstream has a plurality of frames, and each of the frames has a base layer and a plurality of enhancement layers. Next, complexity of the frames is analyzed according to the features of a bitstream parameter of the SVC bitstream. Next, the base layers of the frames are selectively packed to form a base layer packet according to the complexity of the frame. Thereafter, the remainder of the SVC bitstream is packed to form a plurality of enhancement layers according to a packing mechanism of the base layer packet.

Abstract: The present invention provides a system for allocating a bitstream of scalable video coding (SVC). The system includes an adapter and an extractor. The adapter is employed for receiving a bitstream produced by a video encoder and producing an extracting parameter of the bitstream according to the feature of a bitstream parameter in the bitstream content. The extractor is employed for extracting and allocating the bitstream according to the above-mentioned extracting parameter. In this way, the present invention is able to achieve the maximum usage efficiency of bandwidth and the optimum frame quality.

Abstract: A method for storing interpolation data is provided. The method uses a buffer in a cache memory and the concept of memory overlap record for storing previously calculated interpolation data, so as to avoid repeated interpolation, thereby decreasing the amount of system operation and the frequency of reading integer points for calculating interpolation from an external memory. Furthermore, a method of data storage for the buffer is provided. The storage method uses the concept of memory address rotation to store interpolation data beyond the boundary of the buffer. Moreover, another storage method is provided, which distributes interpolation data into a plurality of regions in the buffer according to different combinations of decimal coordinates of the interpolation points for economizing the use of memory space and simplifying the search of interpolation data in the buffer.