Abstract: A video signal encoder/decoder with a 3D noise reduction function and a method thereof. The encoder comprises a storage module, a motion estimation module, a motion compensation module, a first noise reduction module and a coding module. The storage module stores at least one reference image. The motion estimation module receives a first image from an image input end and estimates a motion vector in accordance with the first image and the reference image. The motion compensation module produces motion compensation according to the reference image and the motion vector. The first noise reduction module produces a first noise reduction value with a temporal sequence association according to the first image and the motion compensation. The coding module produces coding data according to the motion compensation and the first noise reduction value.

Abstract: An object image correction apparatus and method for object identification are disclosed. The object image correction method is firstly used for correcting a face or an object under a right position. For example, in order to reduce time consumption for facial identification, the method corrects the deviations such as a rotation, direction, and scaling before an identification process. Preferably, an image is retrieved in a first step. One or more object positions are then detected. Next, some positions of the features are found, and the positions of plural feature points thereon are computed. The method then goes to determine the degree of deviations for the object based on the positions of feature points. Moreover, one or in combination of a rotation correction, a scaling correction, a direction correction, and a shift correction is introduced to process the correction on each deviation. The positions of the feature points are consequently obtained.

Abstract: An image compression method s used for processing a plurality of pixels of an image. The image compression method includes the steps of receiving N successive pixels; obtaining N pixel differences and N map differences corresponding to the N pixels according to values of the N pixels through differential pulse code modulation; calculating to obtain a quantization reference value corresponding to the N pixels according to the N pixel differences; obtaining a quantization parameter corresponding to the N pixels according to the quantization reference value; and encoding the N map differences according to the quantization parameter.

Abstract: A feature data compression device, a multi-directional face detection system and a detection method thereof. The feature data compression device comprises a mirror unit and a transposition unit. The mirror unit compares and computes a plurality of feature values each other to obtain a position of each mutual symmetric feature value and abandoning one of each two symmetric feature values to form first mirrored data, or to obtain a position of the feature value with the highest weight value to form second mirrored data. The transposition unit transposes at least one image template with an angle based on a transposition table for obtaining transposed data.

Abstract: A method and an apparatus for generating a panorama are provided. In the present method, a plurality of raw images of a scene is captured. A coarse motion estimation is executed on the raw images to obtain a coarse motion estimation result of the scene. In the meantime, the raw images inside a window of interest (WOI) are cropped to obtain a plurality of cropped images, and a fine motion estimation is executed on the cropped images according to the coarse motion estimation result of the scene, so as to obtain a fine motion estimation result of the scene. The raw images are stitched and blended according to the fine motion estimation result, so as to generate the panorama.

Abstract: A feature data compression device, a multi-directional face detection system and a detection method thereof. The feature data compression device comprises a mirror unit and a transposition unit. The mirror unit compares and computes a plurality of feature values each other to obtain a position of each mutual symmetric feature value and abandoning one of each two symmetric feature values to form first mirrored data, or to obtain a position of the feature value with the highest weight value to form second mirrored data. The transposition unit transposes at least one image template with an angle based on a transposition table for obtaining transposed data.

Abstract: A method and an apparatus for capturing a three-dimensional image and an apparatus for displaying the three-dimensional image are provided. The method is adapted to an image capturing apparatus including the following steps. First, a plurality of images are captured with continuous pan. Next, a disparity between each two adjacent images is estimated from the images. Then, the images and the disparities are stored.

Abstract: An apparatus capable of detecting location of object contained in image data and its detecting method are disclosed. The apparatus comprises an image capturing module, a weight assignment module, and a processing module. The image capturing module is for capturing an image. The weight assignment module performs the pixel weight/probability assignment according to the priori information and the image, and figures out the initial gravity center of the object according to the object location initialization. The processing module performs the statistical analysis according to the result of the pixel weight/probability assignment and the initial gravity center of the object so as to obtain the analysis result and update the object location. The processing module determines whether or not the analysis result meets the preset value, if it does, the processing module outputs an estimated result; if it doesn't, the processing module repeats the foregoing processes.

Abstract: An automobile equipment control system and a control method thereof. The system comprises an input module, a storage module, an image capturing module, a first recognizing module, a second recognizing module and a processing module. The input module is used to input a plurality of vehicle equipment setting values. The storage module is used to save a plurality of facial characteristic values and a plurality of setting values of vehicle equipment. The image capturing module is used to capture the image of the user. The first recognizing module analyzes the facial image in the image and extracts a facial characteristic point from the facial image. The second recognizing module analyzes the user's hand gesture, the user's head gestures or what the user speaks so as to decide which setting value of vehicle equipment is adopted to control the vehicle equipment.

Abstract: A multidirectional face detection method is for detecting a face in a picture under detection at different positions. The face detection method includes the steps. A selecting window sets to sequentially select different sub-image patterns from the picture under detection. A facial feature weight calculates and it is calculated according to a feature value of the pixels in a sub-image pattern selected by the selecting window, thereby determining if the sub-image pattern has any features similar to the face. A facial edge weight calculates for made on the picture under detection according to a boundary value of the pixels in the sub-image pattern selected by the selecting window, so as to determine if the part of area of the picture under detection has any facial-boundaries. Profile detection is performed to respectively mark the facial-boundaries in the sub-image patterns with arc segments respectively for the sub-image patterns having the facial-boundaries.

Abstract: A method for beautifying a human face in a digital image is adapted to beautify a face area of an input image. The method includes setting a selection window to select a partial image area in the input image; setting a target pixel in the selection window, and setting other pixels as comparison pixels; performing a detail checking process according to a variance between the target pixel; performing a luminance checking process on the target pixel to determine; performing a nonlinear filtering process to filter the target pixel by using a nonlinear filter to generate a filtered value, and providing a mixing ratio to mix the target pixel with the filtered value at the mixing ratio to generate a completed pixel; replacing the original target pixel with the completed pixel; and repeating the above steps until all pixels are completed.

Abstract: A digital image processing device and processing method thereof are provided. The device includes a digital image capturing module, an image enlarging module, an image correcting module, and an image blending module. The digital image capturing module captures a plurality of first resolution images. The image enlarging module enlarges the first resolution images and produces a plurality of second resolution images. The image correcting module selects a target image and produces a plurality of corrected images. The image blending module performs direction gradient operation on each of the pixels of the target and corrected images and produces a plurality of gradient differential values. The image blending module performs a weighting sum operation on each of the pixels of the target and corrected images and produces a third resolution images.

Abstract: An image processing device and the processing method thereof are provided. The device includes an image correcting module, an object-motion detection module and an image blending module. The image correcting module estimates a plurality of local motions of the non-selected images relative to the target image and a plurality of global motions, and performs a plurality of motion corrections to generate a plurality of corrected images. The object-motion detection module judges whether or not the difference between each of the local motions and the corresponding global motion is greater than a threshold value to generate a plurality of object-motion indicators. The image blending module performs an arithmetic operation on each pixel of the target image and each pixel of the corrected images according to the object-motion indicators so as to generate a super-resolution image.

Abstract: An image processing method and an image processing apparatus are provided. In the image processing method, each of a plurality of original images is divided into K bands, where K is a positive integer greater than or equal to 1. A plurality of local motions of each band in any two adjacent of the original images are calculated, and a global motion of each band is calculated according to the local motions of the band. A horizontal component velocity and a vertical component velocity of each band are generated according to the global motion and a capturing time difference. A plurality of transformation matrixes corresponding to the bands are generated according to the horizontal component velocities, the vertical component velocities, a row readout time, and the width of a valid area. One of the two adjacent original images is transformed according to the transformation matrixes to generate a compensated image.

Abstract: A hierarchical face recognition training method and a hierarchical face recognition method thereof for performing a face feature recognition on an image under detection. The method includes a training process and a recognition process. The recognition method includes the steps. A plurality of training samples is obtained. The training samples are subdivided into a plurality of sub-image categories according to a plurality of angle intervals, and the training of a plurality of face features performs on a corresponding sub-image detector of each of the sub-image categories. The training measures performed repeatedly to generate sub-image categories at a sub-level of the sub-image categories. The training method includes the steps. An image under detection is loaded. A similarity of each of sub-image detectors compares according to the image under detection, and the sub-image detector having the highest similarity is selected. The face recognition measures performed repeatedly on the selected sub-image detector.

Abstract: An automatic capturing method of a panoramic image for a digital camera panoramic image is presented. In the method, a motion vector value of an alignment image selected from a first image captured by the digital camera in a real-time image is continuously tracked, calculated, and accumulated to obtain an accumulated motion vector value. When the accumulated motion vector value reaches a preset threshold value, the digital camera is automatically driven to capture a second image, so as to stitch and blend the first image and the second image into a panoramic image. Through the automatic image capture mode, operations for capturing the panoramic image are simplified.

Abstract: A method of establishing a skin color model includes the following steps. A human face detecting procedure is performed on an input image, and a human face area in the input image is circled through a selecting window. A skin color model is established by using a Gaussian probability distribution function (PDF) according to color information in the selecting window. When the skin color model established in the above step is applied to skin color detection, pixels having the skin color in the input image are detected through the skin color model and a Mahalanobis distance computing procedure.

Abstract: An image compression method is used for processing a plurality of pixels of an image. The image compression method includes the steps of receiving N pixels; analyzing the N pixels and generating a content type corresponding to the N pixels; obtaining a quantization parameter and a coding parameter corresponding to the N pixels according to the content type and a currently available buffer space value; obtaining N quantized differences corresponding to the N pixels through a prediction and quantization way according to values of the N pixels and the quantization parameter; and encoding the N quantized differences according to the N quantized differences and the coding parameter.

Abstract: A method for adjusting a skin color of a digital image adjusts the skin color of an input image. The method includes performing a skin color detection process on the input image to generate a skin-color probability plot Sp in a size corresponding to the input image; providing a hue-saturation lookup table named LUT_Color; performing a skin-color reproduction process on the input image to look up the LUT_Color for a chrominance pixel value for each pixel value of the input image to generate a first image, and adjust each pixel value of the first image by using the skin-color probability plot Sp to generate a second image; performing a skin color smoothing process on the second image to generate a third image; and mixing pixel values of the input image and the third image to generate a target image.

Abstract: A method and an apparatus for generating a panorama are provided. In the present method, a plurality of raw images of a scene is captured. A coarse motion estimation is executed on the raw images to obtain a coarse motion estimation result of the scene. In the meantime, the raw images inside a window of interest (WOI) are cropped to obtain a plurality of cropped images, and a fine motion estimation is executed on the cropped images according to the coarse motion estimation result of the scene, so as to obtain a fine motion estimation result of the scene. The raw images are stitched and blended according to the fine motion estimation result, so as to generate the panorama.