Abstract: An image processing apparatus includes a region detection unit, an image processing unit, and an image composing unit. The region detection unit detects a specified region from an original image. The image processing unit deforms a shape of a sample image that is an image serving as a sample in accordance with a shape of the specified region and deforms a texture of the sample image so as to express a depth corresponding to the specified region. The image composing unit overlays the deformed sample image on the specified region of the original image.

Abstract: An encoded bit stream is processed by a lossless decoding unit, an inverse quantization unit, and an inverse orthogonal transform unit in this order, to obtain orthogonally transformed coefficient data and encoding parameter information. The inverse orthogonal transform unit performs an inverse orthogonal transform on the coefficient data by using bases that are set beforehand in accordance with the locations of transform blocks in a macroblock indicated by the encoding parameter information. In this manner, prediction error data is obtained. An intra prediction unit generates predicted image data. An addition unit adds the predicted image data to the prediction error data, to decode image data. By using bases that are set in accordance with the locations of transform blocks, an optimum inverse orthogonal transform can be performed, and encoding efficiency can be increased.

Abstract: Disclosed are techniques for coding coefficients of a video block having a non-square shape defined by a width and a height, comprising coding one or more of x- and y-coordinates that indicate a position of a last non-zero coefficient within the block according to an associated scanning order, including coding each coordinate by determining one or more contexts used to code the coordinate based on one of the width and the height that corresponds to the coordinate, and coding the coordinate by performing a context adaptive entropy coding process based on the contexts. Also disclosed are techniques for coding information that identifies positions of non-zero coefficients within the block, including determining one or more contexts used to code the information based on one or more of the width and the height, and coding the information by performing a context adaptive entropy coding process based on the contexts.

Abstract: A method of irregular motion compensation includes using contours of objects in a reference image to tile the reference image into a plurality of irregular shapes, and mapping each irregular shape to a location in a target image by assigning a motion vector to each irregular shape.