Abstract: A method and system for reducing or eliminating artifacts that may occur during color reduction of images. A bias is added to a data structure that is arranged to store counts of pixel colors of an image. Counts of pixel colors of the image are added to the data structure. Via the data structure, colors from the image are grouped into buckets until the number of buckets is less than or equal to a number of colors in a palette. Then, colors of pixels in the image are mapped to colors in the palette.

Abstract: For operating an operation target object displayed on a screen using a pointing device, a first input coordinate value indicating a position on the screen designated by the pointing device and a second input coordinate value detected discretely from, and before, the first input coordinate value are detected. A changing direction is calculated based on the first input coordinate value and the second input coordinate value, and a rotation direction is determined based on the changing direction. The operation target object is rotated in the rotation direction.

Abstract: A method, device and computer system for creating a smooth, continuous height (scalar or vector) field are described. The described techniques permit arbitrary closed regions to be smoothly shaded without producing unnatural smoothness at the region's edges or boundaries.

Abstract: An imaging or other sensory reproduction system efficiently converts image or other sensory data between a perceptual color space (e.g., the sRGB color space) and a physical color space (unity gamma) or other perceptual/physical sensory models that are related by an expression involving a computationally expensive exponential function. The imaging system calculates exponential functions that can be composed from computationally inexpensive operations, such as square root, square, reciprocal, as well as multiplications and/or additions and subtractions. These computationally less expensive functions are then combined, such as in a weighted and/or offset mean, summation or difference to approximate the computationally expensive exponential function. The imaging system evaluates the expression using the approximation to efficiently yield the converted image data.

Abstract: A system and method are provided for computing partial differential equations in a hardware graphics pipeline. Initially, input is received in a hardware graphics pipeline. Next, the input is processed to generate a solution to a partial differential equation utilizing the hardware graphics pipeline.

Abstract: This is an image processing device for predicting the lower-order bits of target pixel data, based on one or more pieces of pixel data constituting image data. The image processing device comprises a lower-order bit calculation unit for calculating the lower-order bits of the target pixel data, based on one or more pieces of pixel data constituting image data and specifying the data as corrected lower-order bits and a lower-order bit superimposition unit for superimposing the corrected lower-order bits calculated by the lower-order bit calculation unit on higher-order bits of the target pixel data.

Abstract: A graphics engine includes a setup unit and a rendering unit. The setup unit computes coefficients A, B, and C used for interpolating an attribute v of a triangle to be rendered for a graphics image. The setup unit then derives compressed coefficients Ã, {tilde over (B)}, and {tilde over (C)} based on the coefficients A, B, and C. The compressed coefficients have a fixed-point format with R integer bits left of a binary point and T fractional bits right of the binary point, where R>1 and T?0. R is selected based on the number of bits used for attribute v, T is selected based on the screen dimension, and R+T is much less than the number of bits used to represent the coefficients A, B, and C. The rendering unit performs interpolation for the attribute v using the compressed coefficients Ã, {tilde over (B)}, and {tilde over (C)}, and may be implemented with a simple (R+T)-bit non-saturating accumulator.

Abstract: A method is for calculating a height of a chuck top. A height of the top surface of the chuck top which corresponds to an arbitrary position specified on the XY coordinate plane by a computer is calculated in each of the four quadrants based on a coordinate transformation formulas. The method includes setting, by using the computer, a conical model in which two adjacent points other than the center point of the chuck top which correspond to the specified coordinates in a predetermined quadrant of the XY coordinate plane are obtained on a circumference having the center point of the chuck top as the origin and specifying an arbitrary point in the predetermined quadrant by using the computer and calculating a height of the arbitrary point of the chuck top based on the conical model, the coordinate transformation formulas and the specified coordinates.

Abstract: A technique for graphically processing perimeter information includes determining a polar representation of the perimeter of an object to define polar perimeter information, storing the polar perimeter information, retrieving the stored polar perimeter information using an overlay, and applying the retrieved polar perimeter information to provide a graphical effect within a scene (such as, but not limited to, a shadow effect).

Abstract: A display controller includes a video signal-analyzing unit, a video signal-adjusting unit operable to receive a video signal, to adjust the video signal in accordance with adjustment parameter information, and to feed the adjusted video signal into a display device, and a light source-controlling unit operable to feed a light source-controlling signal into a light source in accordance with light source light-emitting amount information. The video signal-analyzing unit allows timing in which the display device displays a picture in accordance with the adjusted video signal from the video signal-adjusting unit to be synchronized with timing in which the light source changes a light-emitting amount in response to the light source-controlling signal from the light source-controlling unit.

Abstract: In a method for reversibly transforming a data format, a forward transformation and a backward transformation are reciprocally conducted for data between unit systems having different resolution levels, and the forward transformation and the backward transformation, a first unit system having a lower resolution level is used as a common unit system, and a reversible data conversion is conducted by an integer operation for data in the first unit system having the lower resolution level and data in a second unit system having a higher resolution level higher than the first unit system.

Abstract: A method is disclosed for culling an object database in a graphics processing system. In one embodiment, the method comprises encoding per-object parameters and culling parameters. The per-object parameters are encoded in texture format thereby creating at least one per-object texture containing the encoded per-object parameters. Next, a fragment program used in a fragment processor of the GPU is optionally updated. The updated fragment program embodies a culling operation. A polygon is then rendered, wherein the rendering step includes per-fragment operations. During the per-fragment operations, the updated fragment program is executed. The culling operation embodied therein (i) accesses the culling parameter, (ii) samples the per-object textures, and (iii) produces cull results for a set of database objects. In this fashion, the fragment processor in the GPU is leveraged to perform computationally intensive culling operations.

Abstract: Systems and methods are disclosed that facilitate rapidly warping a two-dimensional image using integer math. A warping table can contain two-dimensional floating point output pixel offset values that are mapped to respective input pixel locations in a captured image. The warping table values can be pre-converted to integer offset values and integer grid values mapped to a sub-pixel grid. During warping, each output pixel can be looked up via its integer offset value, and a one-dimensional table lookup for each pixel can be performed to interpolate pixel data based at least in part on the integer grid value of the pixel. Due to the small size of the lookup tables, lookups can potentially be stored in and retrieved from a CPU cache, which stores most recent instructions to facilitate extremely rapid warping and fast table lookups.

Abstract: Generating an image includes receiving a light at sensors, where the light is associated with images. A previous matrix is determined, where the previous matrix includes image information associated with an image. For each sensor, a current image data corresponding to a current image is generated and a current matrix is determined using the previous matrix and the current image data. The current matrix includes image information associated with the current image. A fusion matrix is computed according to the current matrix of each sensor, where the fusion matrix initiates generation of a fused image.

Abstract: A method for transforming an original image to a new image is provided. The original image includes M rows of original data; the new image includes Q rows of new data. The method first generates a (2i?1)th row and a (2i)th row of intermediate data respectively based on the (2i?1)th row and the (2i)th row of original data. Then, the method generates a (2i+1)th row and a (2i+2)th row of intermediate data respectively based on the (2i+1)th row and the (2i+2)th row of original data. During the process of generating the (2i+1)th row of intermediate data, the (2j?1)th row of new data is simultaneously generated based on the (2i?1)th row and the (2i+1)th row of intermediate data. During the process of generating the (2i+2)th row of intermediate data, the (2j)th row of new data is simultaneously generated based on the (2i)th row and the (2i+2)th row of intermediate data.

Abstract: A computer-readable medium of one embodiment of the invention is disclosed that has a data structure stored thereon to represent extended color gamut information for a pixel. The data structure has a number of data fields, where each data field is to store a number of bits representing a different color component of the pixel. The data structure also has at least one interpretation bit for each data field to indicate how the bits of the data field are to be interpreted in representing the different color component of the pixel.

Abstract: A displaying method that makes the following possible: Executing an operation from either of two approaches, i.e. a processing oriented to an object to be processed or a processing oriented to a task to be processed, and effectively dealing with a lot of tasks and various types of objects, and performing a flexible operation. In the displaying method in an information processing apparatus, both displays, i.e. a display intended for selecting the task to be processed and a display intended for selecting the object to be processed, are displayed on a display apparatus in such a manner that the two displays are distributed into two tabs, respectively. This condition makes it possible to select the task to be processed with a selection of the tabs, and also makes it possible to execute in an arbitrary order the processing oriented to the task and the processing oriented to the object.

Abstract: Methods for choosing and combining colors from a color palette to render an image color tone are disclosed. A set of up to four palette colors are chosen and the weighted factors for combining the chosen palette to render the image color are determined. The weighted factors of the chosen palette colors are ordered according to an ordering criterion or criteria. The color output of a display pixel is the chosen palette color associated with the interval in which the threshold value falls. Color data compression may also be achieved by eliminating at least one color from the set of chosen palette colors used to render an image color that fails to exceed a specified threshold value. Also disclosed are methods for designing uniform and non-uniform color palettes.

Abstract: In a color correcting parameter calculator, a color correcting unit 11 calculates a target color by executing color correction processing of an input color on a reference side while a reference side color perception value calculator unit 13 calculates color perception values of the input color and its target color on the reference side. An object side color perception value calculator unit 15 calculates color perception values of an input color on an object side. A color difference searching unit 16 detects that input color on the reference side which has a minimum color difference with the input color on the object side, by accessing a color perception storage unit 14 for input and target colors for the reference side. A device dependent color unit 17 on the object side calculates device dependent color data of a target color on the object side with reference to perception values of the target color on the reference side while a color correcting parameter calculator 18 calculates color correcting parameters.

Abstract: Several embodiments of the present application disclose techniques, systems and methods for changing or rendering input image data that may assume a first white point for a given display into image data to be rendered under a second—assumed, desired or measured—white point of the display.

Abstract: An orthogonal data converter for converting the components of a sequential vector component flow to a parallel vector component flow. The data converter has an input rotator configured to rotate corresponding vector components of the sequential vector component flow by a prescribed amount, and a bank of register files configured to store the rotated vector components. The converter also has an output rotator configured to rotate the position of the vector components read from the bank of register files by a prescribed amount. A controller of the converter is operative to control the addressing of the bank of register files and the rotating of the vector components. In this regard, the controller is operative to write the vector components to the bank of register files in a prescribed order and read the vector components in a prescribed order to generate the parallel vector component flow.

Abstract: A technique for graphically processing perimeter information includes determining a polar representation of the perimeter of an object to define polar perimeter information, storing the polar perimeter information, retrieving the stored polar perimeter information using an overlay, and applying the retrieved polar perimeter information to provide a graphical effect within a scene (such as, but not limited to, a shadow effect).

Abstract: A multi-chip system is disclosed for distributing the convolution process. Rather than having multiple convolution chips working in parallel with each chip working on a different portion of the screen, a new design utilizes chips working in series. Each chip is responsible for a different interleaved region of screen space. Each chip performs part of the convolution process for a pixel and sends a partial result on to the next chip. The final chip completes the convolution and stores the filtered pixel. An alternate design interconnects chips in groups. The chips within a group operate in series, whereas the groups may operate in parallel.

Abstract: According to one embodiment, a method for compensating for inadequate bit resolution in a light processing system includes receiving a plurality of values each indicative of an intensity level for a pixel to be displayed. Each of the values is represented by a plurality of bits of data. The method also includes determining a quantization step size for the plurality of bits of data. For at least one particular pixel of the pixels, a set of consecutive pixels including the particular pixel is selected. The method also includes determining a difference between the value associated with the particular pixel in the set and each value associated with the other pixels in the set, and also determining that all of the determined differences are less than or equal to the quantization step size. In response, a filtered value for the particular pixel in the set is generated based at least on some of the pixels in the set in addition to the particular pixel.

Abstract: A display control device is provided which includes characteristic value-calculating unit for calculating a characteristic value based on an input image signal, a conversion characteristic-calculating unit for determining a conversion characteristic from the characteristic value, and a signal-converting unit for converting the input image signals in accordance with the determined conversion characteristic. The conversion characteristic has a low level region segment, which is close to the origin of axes respectively representing the input image signal and an output image signal of the display control device, a high level region segment, which is close to a full scale point, and a middle level region segment, which is positioned between the low level region and high level region, and the slope in the middle level region is set so as to be larger than both of the slopes of the low level region and the high level region.

Abstract: A method, device and computer system for creating a smooth, continuous height (scalar or vector) field are described. The described techniques permit arbitrary closed regions to be smoothly shaded without producing unnatural smoothness at the region's edges or boundaries.

Abstract: The gamma adjustment allows the luminance for the sub-pixel arrangement to match the non-linear gamma response of the human eye's luminance channel, while the chrominance can match the linear response of the human eye's chrominance channels. The gamma correction allows the algorithms to operate independently of the actual gamma of a display device. The sub-pixel rendering techniques disclosed with gamma adjustment can be optimized for a display device gamma to improve response time, dot inversion balance, and contrast because gamma correction and compensation of the sub-pixel rendering algorithm provides the desired gamma through sub-pixel rendering. These techniques can adhere to any specified gamma transfer curve.

Abstract: A computer method, apparatus and storage medium is provided for creating quantitative aesthetic graphics from data. The invention utilizes a graph algebra to construct graphs and visually or otherwise represents the graphs as a quantitative aesthetic graphic representation. To create the quantitative aesthetic graphics from data, the data is indexed to form a data set. Thereafter, the data is converted into a variable data structure composed of an index set, a range and a function. The variable data structure is converted into a variable set by using at least one of a blend step, a cross step and a nest step. The variable set is mapped into a set of points and the set of points is mapped into an aesthetic representation.

Abstract: The systems detect movement of text or areas of high spatial frequency in one frame to another frame. If such movement is detected and meets a certain level or threshold, the subpixel rendering processing of such text or areas of high spatial frequency can be changed.

Abstract: A pixel interpolation method includes: utilizing a plurality of pixel pairs on two successive scan lines to determine at least one first possible angle; utilizing a plurality of pixel value distribution trends of the two successive scan lines to determine at least one second possible angle; comparing pixel values of a plurality of pairs of neighboring pixels to detect whether the pixel values are larger than a threshold value in order to determine at least one third possible angle; determine a most appropriate angle according to the first, second, and third possible angles; detecting whether the most appropriate angle is correct by examining that the most appropriate angle is in a group; and utilizing a first pixel on the upper scan line of the two successive scan lines and a second pixel on the bottom scan line of the two successive scan lines to interpolate the a target pixel.

Abstract: A liquid crystal display is provided, which includes: a liquid crystal panel assembly including a plurality of pixels connected to a plurality of gate lines and a plurality of data lines; a signal controller for processing image data, the signal controller including a dynamic capacitance capture (“DCC”) block for modifying image data assigned to the pixels by selectively performing DCC on the image data based on the difference between the image data of a current frame (“current data”) and the image data of a previous frame (“previous data”); a gate driver for sequentially applying a gate-on voltage to the gate lines of the liquid crystal panel assembly; and a data driver selecting data voltages among a plurality of gray voltages in response to the modified image data from the signal controller and applies the data voltages to the data lines of the liquid crystal panel assembly.

Abstract: A de-interlacing methodology generates frames from interlaced video signals by incorporating data from multiple fields into an interpolation-based de-interlacing process. Pixels directly above and below a blank pixel location and pixels immediately before and after the blank pixel location (in the fields immediately preceding and following, respectively, the blank pixel field) can be used to interpolate a pixel value for the blank pixel location. The use of pixel data from multiple fields improves the resolution of the interpolation process, thereby improving output frame accuracy. Adjacent pixel values can also be adjusted to further improve the consistency of the visual display provided by the output frames.

Abstract: A software-implemented graphics rendering system and method designed and optimized for embedded devices (such as mobile computing devices) using fixed-point operations including a variable-length fixed point representation for numbers and a normalized homogenous coordinates system for vector operations. The graphics rendering system and method includes a fixed-point mathematics library and graphics functions that includes optimized basic functions such as addition, subtraction, multiplication, division, all vertex operations, matrix operations, transform functions and lighting functions, and graphics functions. The mathematical library and graphics functions are modified and optimized by using a variable-length fixed-point representation and a normalized homogenous coordinate system (NHCS) for vector operations.

Abstract: A graphics processing device for converting coefficients in a video data stream from a first type, e.g., frequency-domain, to a second type, e.g., color-domain. The device includes an input for receiving the video data stream including a set of coefficients of the first type and a storage medium holding a data structure containing a first set of coefficients of the second type. The device further includes a processor communicating with the input and with the storage medium. The processor uses the data structure to convert the set of coefficients of the first type to a second set of coefficients of the second type. The device also includes an output in communication with said processor, for releasing an output video data stream including the second set of coefficients of the second type. The same data structure is used repeatedly for each incoming set of coefficients of the first type, thus allowing a transform, such as an IDCT, to be computed efficiently.

Abstract: A hardware-based circuitry for digital processing for color saturation control circuitry, brightness control circuitry, contrast control circuitry, and color hue control circuitry, some or all of which can be added to the luminance/chrominance (Y/C) signals to the red green blue (RGB) digital YCbCr-to-RGB conversion circuitry that is required by devices such as an LCoS display device. (these controls can also be used without the Y/C-to-RGB conversion surrounding. Important is the fact that the signals for control be in Y/C domain) A digital video signal from the source in the Y/C (luminance/chrominance) domain is fed into processing circuitry, where the C (chrominance) component is split into Cb and Cr subcomponents. Digital YCbCr-to-RGB conversion circuitry transforms the Y/C signals to RGB domain, required by, for example, the LCoS display device.

Abstract: An image transformation method for translating a non-linear 2D geometrical transformation into two separable 1D geometrical transformations first determines the inverse of the 2D geometrical transformation to form an inverse 2D geometrical transformation. Then the method converts the inverse 2D geometrical transformation into an analytical inverted 2D geometrical transformation and separates the analytical inverse 2D geometrical transformation into first and second 1D geometrical transformations. The method then represents said inverse 2D geometrical transformation and first and second 1D geometrical transformations as tensor spline surfaces and then compares an evaluation of said first and second 1D geometrical transformations at each pixel with an evaluation of the analytical inverse 2D geometrical transformation at each pixel. If the error evaluation does not meet a predetermined level of performance then the separation and transformation steps are repeated.

Abstract: An n-filter according to the present invention includes a nonlinear filter, a pattern detector, and a switch. The nonlinear filter maintains a steep edge whose size is larger than a predetermined threshold in fluctuations of pixels constituting an input image signal and, at the same time, smoothes a non-edge portion that does not include the edge. The pattern detector detects a fine edge in the fluctuations of the pixels constituting the input image signal and notifies the switch that the fine edge exists. The switch outputs the input image signal or an image signal output from the nonlinear filter to the subsequent stage in accordance with the notification from the pattern detector.

Abstract: The present method represents a three-dimensional shape model by polygons according to a plurality of object images information picked up by rotating a real object for every arbitrary angle to assign texture information on each polygon from object image information having the largest projection area of the relevant polygon. In order to improve the color continuity between adjacent polygons, the object image information having correspondence between a polygon of interest and an adjacent polygon thereof is selected so as to be the object image information approximating the shooting position and the shooting direction. An alternative method divides an object image into a plurality of regions, obtains difference between an object image and a background image in region level, outputs a mean value of the absolute value of difference in the region level, and detects the region having the mean value of absolute values of difference equal to or greater than a threshold value as the object portion.

Abstract: A user gives, as input, two curves at a start time and at an end time and reference correspondence points with respect to the respective curves. Then, a pursuit operation of pursuing a reference correspondence point is carried out with respect to picture images successive in the time axis direction. As a result, positions of reference correspondence points at respective intermediate stages can be determined. Then, interpolation of the shape is carried out. Then, the determined interpolated shape is deformed in correspondence with the determined reference correspondence point. Accordingly, even in the case where a contour shape does not linearly move, it is possible to precisely pursue correspondence points on the contour. Thus, a contour shape with higher precision can be prepared.

Abstract: The present invention provides methods, apparatus and systems for enabling real-time lighting of 3D geometric models. While these are well-suited for all processors, these are particularly useful on low power processors typically found in PDAs, cell phones, and embedded display systems. Methods+presented utilize a preprocessing scale factor and real-time shift factor to efficiently calculate diffuse lighting intensities for the bound vertices of the geometric model. An example embodiment includes scaling first digitized data of a 3D model in a floating point datatype by a scaling factor to form scaled digitized data. The scaling factor being a proper power of two. Converting said scaled digitized data into an integer datatype forming second digitized data. Enabling a transfer of said second digitized data to a display system, and facilitating lighting and shading said second format of digitized data.

Abstract: An area within a frame is defined in which text or an image is to be generated. The text or image may also include an optional border surrounding and adjacent to the text or image. In accordance with one embodiment, in order to increase the contrast between the text or image and the background area of the frame, one or more color component values are selected for use in generating the text/image (or associated border). Specifically, a complement of one or more color component values associated with one or more pixels of the frame are determined. The complement may, for example, be obtained from an average color component value over a set of pixels. Alternatively, the complement may be obtained directly for each of a set of pixels in the frame from which an average value may then be obtained. The complemented-averaged color component value(s) may then be used to generate the text/image and/or associated border.

Abstract: A data conversion circuit is provided for displaying a high definition image whose resolution is different from that of a display screen having a cell arrangement that is not a square arrangement. The data conversion circuit performs addition operation with weighting that is also resolution conversion of the integer ratio M:N and data correction for improving a linear display quality for input image data.

Abstract: In an embodiment, a functional unit including a compressor section and a 36-bit SIMD adder is used to perform a STMD four-pixel averaging instruction. The functional unit generates four four-pixel averages. Four pixel values and a rounding value are compressed into a sum and a carry vector. The two least significant bits of the sum vector and the LSB of the carry vector are dropped before being input to the 36-bit SIMD adder. The two resultant 8-bit vectors are added by the 36-bit adder to directly generate the average pixel value result.

Abstract: In an information processing apparatus for conducting an affine transformation representative of ( x ? y ? z ? ) = A ? ( x y z ) + ( t 1 t 2 t 3 ) , ? a ? ? matrix A ? = ( a 11 ? a 12 ? a 13 ? a 21 ? a 22 ? a 23 ? a 31 ? a 32 ? a 33 ? ) = ? ? ? A = ( ? ? ? a 11 ? ? ? a 12 ? ? ? a 13 ? ? ? a 21 ? ? ? a 22 ? ? ? a 23 ? ? ? a 31 ? ? ? a 32 ? ? ? a 33 ) obtained by multiplying a matrix A for the affine transformation by ?(?0) is stored in a memory section in advance.

Abstract: A computer method, apparatus and storage medium is provided for creating quantitative aesthetic graphics from data. The invention utilizes a graph algebra to construct graphs and visually or otherwise represents the graphs as a quantitative aesthetic graphic representation. To create the quantitative aesthetic graphics from data, the data is indexed to form a data set. Thereafter, the data is converted into a variable data structure composed of an index set, a range and a function. The variable data structure is converted into a variable set by using at least one of a blend step, a cross step and a nest step. The variable set is mapped into a set of points and the set of points is mapped into an aesthetic representation.

Abstract: A brightness control apparatus comprising a probability density function (PDF) calculator to calculate a PDF based on pixel values of respective pixels of an input image signal, a first setter to set an upper limit value and a lower limit value with respect to the pixel values that are equal to or smaller than a predetermined level in the PDF, and a brightness value controller to calculate a cumulative distribution function for the PDF controlled by the upper and lower limit values set by the first setter, and calculating brightness levels corresponding to the input image signal based on the calculated cumulative distribution function. The brightness control apparatus controls the brightness of an image signal so as to prevent it from becoming unnecessarily bright or dark when compensating the brightness of an image signal, while also preventing degradation of contrast.

Abstract: The present invention is directed to a method of discriminating between textual content and graphical content. The method includes the steps of receiving a plurality of pixel values for a pixel line segment, calculating a plurality of spatial gradients based on the pixel values of adjacent pixels, determining a smoothness index by processing the plurality of spatial gradients, and identifying the pixel line segment as text or graphics by comparing the smoothness index to a threshold value.

Abstract: An image processing apparatus is provided for accurately recognizing an edge direction to perform an accurate image interpolation. A direction determining unit recognizes an edge direction of a remarked pixel and outputs it with information on its position to a reliability ranking unit and a directional distribution generating unit. A direction interpolating unit interpolates the remarked pixel in terms of directional interpolation. The reliability ranking unit determines whether or not a interpolated pixel is properly interpolated by the direction interpolating unit, ranks its reliability, and outputs a result to a directional distribution generating unit. This directional distribution generating unit generates directional distribution based on directional information and reliability information. A direction selecting unit recognizes an edge direction based on the directional distribution generated by the directional distribution generating unit.

Abstract: The image coding device comprises data adding means for adding specific data to input image data at the end of image data, and arithmetic coding unit not issuing remaining output code of code register after coding of final input data. In this constitution, increase of circuit scale can be suppressed and decline of operation clock can be prevented.

Abstract: An image processing apparatus is provided for accurately recognizing an edge direction to perform an accurate image interpolation. A direction determining unit recognizes an edge direction of a remarked pixel and outputs it with information on its position to a reliability ranking unit and a directional distribution generating unit. A direction interpolating unit interpolates the remarked pixel in terms of directional interpolation. The reliability ranking unit determines whether or not a interpolated pixel is properly interpolated by the direction interpolating unit, ranks its reliability, and outputs a result to a directional distribution generating unit. This directional distribution generating unit generates directional distribution based on directional information and reliability information. A direction selecting unit recognizes an edge direction based on the directional distribution generated by the directional distribution generating unit.