Abstract: An apparatus is provided for detecting blur in an image. The apparatus includes an image generator that is configured to generate a plurality of corresponding different resolution images based on a base image. The plurality of corresponding different resolution images is provided to an edge detector. The edge detector detects edge transitions in each of the plurality of corresponding different resolution images and provides edge maps to an edge parameter comparator. The edge parameter comparator compares corresponding edge parameters as detected by the edge detector and provides a result map to a blur calculator. The blur calculator determines at least one blur parameter based on the result map and provides the blur parameter to a blur detector. The blur detector then determines if the base image is blurred based on a comparison of the blur parameter with at least one blur parameter threshold.

Abstract: Red-eye detection based on red region detection with eye confirmation initially identifies pixels that correspond to the color of red-eye within an image. A determination is then made as to whether these identified pixels and surrounding areas are part of an eye or not part of an eye. Those identified pixels that are determined to be part of an eye are the detected red-eye regions.

Abstract: A system and method that includes a virtual human face generation technique which synthesizes images of a human face at a variety of poses. This is preferably accomplished using just a frontal and profile image of a specific subject. An automatic deformation technique is used to align the features of a generic 3-D graphic face model with the corresponding features of these pre-provided images of the subject. Specifically, a generic frontal face model is aligned with the frontal image and a generic profile face model is aligned with the profile image. The deformation procedure results in a single 3-D face model of the specific human face. It precisely reflects the geometric features of the specific subject. After that, subdivision spline surface construction and multi-direction texture mapping techniques are used to smooth the model and endow photometric detail to the specific 3-D geometric face model.

Abstract: An image processing system for automatically extracting image frames suitable for printing and/or visual presentation from a compressed image data is described. The image processing system includes a face detector that detects if an image frame contains at least a face. The image processing system also includes a blur detector that determines the blur indicator value of the image frame directly using the information contained in the compressed image data if the image frame is determined to contain a face. The blur detector indicates that the image frame is suitable for printing and/or visual presentation if the blur indicator value of the image frame is less than a predetermined threshold. The image processing system may also include a motion analyzer that determines if the image frame is a super-resolution image frame suitable for printing and/or visual presentation if the image frame does not contain any face.

Abstract: A texture-based text localization system proceeds directly in the compressed domain for DCT compressed JPEG images or MPEG videos. The DCT coefficient values in JPEG images and MPEG videos, which capture the directionality and periodicity of local image blocks, are used as texture feature measures to classify text areas. Each unit block in the compressed images is classified as either text or nontext. In addition, post-processing in both the compressed domain and the reconstructed candidate text areas can be used to refine the results. For video frames that contain text, the displacement of text between two consecutive frames is estimated which gives the velocity of the moving text. This temporal displacement information is also used to further refine the localization results. The text is then processed to provide content or speech output.