Abstract: Gesture recognition methods and systems are provided. First, a plurality of gesture templates are provided, wherein each gesture template defines a first gesture characteristic and a corresponding specific gesture. Then, a plurality of images is obtained, and a multi-background model is generated accordingly. At least one object image is obtained according to the multi-background model, wherein the object image includes at least an object having a plurality of edges. The included angles of any two adjacent edges of the object image are gathered as statistics to obtain a second gesture characteristic corresponding to the object image. The second gesture characteristic of the object image is compared with the first gesture characteristic of each gesture template. The specific gesture corresponding to the first gesture characteristic is obtained, when the second gesture characteristic is similar to the first gesture characteristic.

Abstract: A gesture recognition system includes an image pick-up device, a processor, an operation engine, an optimal template selection means, and a display terminal. The image pick-up device is for capturing an image containing a natural gesture. The processor is for finding out a skin edge of a skin part from the image, and then classifying the skin edge into multiple edge parts at different angles. The operation engine has multiple parallel operation units and multiple gesture template libraries of different angle classes. These parallel operation units respectively find out gesture templates most resembling the edge parts in the gesture template libraries of different angle classes. The optimal template selection means selects an optimal gesture template from the resembling gesture templates found out by the parallel operation units. The display terminal is for displaying an image of the optimal gesture template. Thereby, marker-less and real-time gesture recognition is achieved.

Abstract: A method for producing an image with depth by using 2D image includes obtaining a set of internal parameters of a camera. The camera takes at least a first and a second 2D images with a small shift. The first 2D image has N depths, and N?2. Several sets of external parameters of the camera corresponding to the 2D images are estimated. A 3D information respectively corresponding to the N depths of the first 2D image at each pixel or block is calculated. A proper depth of each pixel or image block is determined. Through the internal parameters, the external parameters, and the N depths, each pixel or image block of the first 2D image is projected onto N positions of the second 2D image, so as to perform a matching comparison analysis with the second 2D image, thereby determining the proper depth from the N depths.

Abstract: Image filling methods. A plurality of images corresponding to a target object or a scene are captured at various angles. An epipolar geometry relationship between a filling source image and a specific image within the images is calculated. The filling source image and the specific image are rectified according to the epipolar geometry relationship. At least one filling target area in the rectified specific image is patched according to the rectified filling source image.

Abstract: Gesture recognition methods and systems are provided. First, a plurality of gesture templates are provided, wherein each gesture template defines a first gesture characteristic and a corresponding specific gesture. Then, a plurality of images is obtained, and a multi-background model is generated accordingly. At least one object image is obtained according to the multi-background model, wherein the object image includes at least an object having a plurality of edges. The included angles of any two adjacent edges of the object image are gathered as statistics to obtain a second gesture characteristic corresponding to the object image. The second gesture characteristic of the object image is compared with the first gesture characteristic of each gesture template. The specific gesture corresponding to the first gesture characteristic is obtained, when the second gesture characteristic is similar to the first gesture characteristic.

Abstract: A gesture recognition system includes an image pick-up device, a processor, an operation engine, an optimal template selection means, and a display terminal. The image pick-up device is for capturing an image containing a natural gesture. The processor is for finding out a skin edge of a skin part from the image, and then classifying the skin edge into multiple edge parts at different angles. The operation engine has multiple parallel operation units and multiple gesture template libraries of different angle classes. These parallel operation units respectively find out gesture templates most resembling the edge parts in the gesture template libraries of different angle classes. The optimal template selection means selects an optimal gesture template from the resembling gesture templates found out by the parallel operation units. The display terminal is for displaying an image of the optimal gesture template. Thereby, marker-less and real-time gesture recognition is achieved.

Abstract: A method for rapidly building image space relation using plane filtering limitation is provided. The method comprises first getting a plurality of continuous image data taken from different shooting angles. Then, a plurality of feature points of the image data is initialized. Next, a number of feature points of those image data is randomly extracted and compared to determine whether the compared feature points are coplanar or not. When the selected feature points are not coplanar, an epipolar fundamental matrix is calculated according to the selected feature points. On the contrary, when the selected feature points are coplanar, a medium error value for all the feature points is set as a maximum value.

Abstract: The present invention provides a multi-touch position tracking technique and an interactive system and a multi-touch interactive image processing method using the same. In the present invention, a light guide element is designed to comprise frustrating structures to frustrate total internal reflection (TIR) so that the light beam therein can be dispersed to form a dispersed optical field distribution. The dispersed optical field is used to respond a physical relation between an object and the light guide element.

Abstract: A method for producing an image with depth by using 2D image includes obtaining a set of internal parameters of a camera. The camera takes at least a first and a second 2D images with a small shift. The first 2D image has N depths, and N?2. Several sets of external parameters of the camera corresponding to the 2D images are estimated. A 3D information respectively corresponding to the N depths of the first 2D image at each pixel or block is calculated. A proper depth of each pixel or image block is determined. Through the internal parameters, the external parameters, and the N depths, each pixel or image block of the first 2D image is projected onto N positions of the second 2D image, so as to perform a matching comparison analysis with the second 2D image, thereby determining the proper depth from the N depths.

Abstract: Image filling methods. A plurality of images corresponding to a target object or a scene are captured at various angles. An epipolar geometry relationship between a filling source image and a specific image within the images is calculated. The filling source image and the specific image are rectified according to the epipolar geometry relationship. At least one filling target area in the rectified specific image is patched according to the rectified filling source image.

Abstract: A method for calibrating a response curve of a camera is provided. A homography relationship of an image sequence captured by the camera is calculated using a coplanar information including feature correspondence blocks of the image sequence. An intensity mapping function is then obtained from the intensity information of the correspondence blocks according to the homography relationship. The calculation for obtaining the intensity mapping function is significantly reduced by focusing on the correspondence blocks, which can also avoid the problem of outliers.

Abstract: A method for rapidly building image space relation using plane filtering limitation is provided. The method comprises first getting a plurality of continuous image data taken from different shooting angles. Then, a plurality of feature points of the image data is initialized. Next, a number of feature points of those image data is randomly extracted and compared to determine whether the compared feature points are coplanar or not. When the selected feature points are not coplanar, an epipolar fundamental matrix is calculated according to the selected feature points. On the contrary, when the selected feature points are coplanar, a medium error value for all the feature points is set as a maximum value.