Abstract: There is described a method for detecting the presence of skin tone in an image. A gray scale representation of a pixel within the image is provided. Next, a red chrominance independent representation for is provided for the pixel. Then, the two representations are analysed to determine whether a difference in value between the representations corresponds with a the presence of a skin tone value. The present invention provides a rapid skin tone detection classifier particularly useful for real time applications.

Abstract: A region where a detecting target object exists is extracted by a comparison between an evaluated value indicating a probability that the detecting target object exists and a threshold, through the process of producing a differential image between different frames in plural frames constituting a continuous image, of setting an average value in an averaging region extended around each pixel of the differential image as a new value of each pixel, of obtaining the evaluated value by applying a filter that acts on a search region on an image to a search region extended around a search pixel that is extracted by comparing the new value and the threshold, on the differential image.

Abstract: A method and apparatus to build a 3-dimensional grid map and a method and apparatus to control an automatic traveling apparatus using the same. In building a 3-dimensional map to discern a current location and a peripheral environment of an unmanned vehicle or a mobile robot, 2-dimensional localization and 3-dimensional image restoration are appropriately used to accurately build the 3-dimensional grid map more rapidly.

Abstract: A program for uploading images or direct scans of fingerprints into a computer system where they are then input into proper forms used in searching common databases. The system develops and saves files into types that are typically used in the biometric search industry, and these files are submitted to national fingerprint databases. This entire process is streamlined and reduces the amount of work and input required by the end user to a minimum. The program will reduce the user's amount of work necessary to submit a fingerprint search to national databases by simplifying the formatting of the submittal page for such national databases, resulting in a more efficient manner of preparing fingerprint search documents in appropriate formats. The program is fully customizable and can be used to satisfy the requirements of any national fingerprint database.

Abstract: A face recognition method includes the step of: receiving feature data acquired by detecting each part of a face of a person included in digital data from a terminal at a server; comparing the feature data received from the terminal with those of persons determined as accessible among all feature data recorded in a feature data database based on face recognition technology, and deciding top n persons with top n degree of similarities among all persons recorded in the feature data database at the server, wherein the feature data recorded in the feature data database include feature data of each part of faces of the persons; and providing the information on the top n persons to the terminal. It may reduce work capacity and processing time by extracting and transmitting to the server feature data of each part of the face such as eyes, nose, mouth, ears and jaws.

Abstract: A method for quality-aware transcoding of an input image into an output image for display on a terminal having device file size and image size constraints are disclosed. The method extracts features of the input image including dimensions and file size of the input image, predicts a file size of the output image taking into account constraints of the terminal and extracted features, including selecting a set of feasible transcoding parameters so that a corresponding predicted file size of the output image meets the device file size constraint of the terminal, determines quality metric values of the output image, characterizing a measure of distortion of the input image introduced by the transcoding, corresponding to the feasible transcoding parameters in the set of feasible transcoding parameters, and selects those transcoding parameters from the set of feasible transcoding parameters, which yield the highest quality metric value. A corresponding system is also provided.

Abstract: A method performed by a software process executing on a computer system, includes obtaining a digital image having a plurality of pixels encoded in a YUV color space. Each pixel has a luma component of value Y, a blue color-difference component of value U and a red color-difference component of value V. For a specified pixel, the method includes calculating whether U is less than a first threshold and V is greater than a second threshold. The method further includes determining whether the specified pixel potentially depicts an orange hue depending on a result of the calculation.

Abstract: An image editing apparatus includes an image input unit including an image pickup circuit which takes a picture of a subject to obtain an image of a subject. A face detection circuit detects an image of a face of the subject from the image obtained from the image input unit. A face expression detection circuit detects at least any one of direction and expression of the face of the subject, based on the image of the face detected by the face detection circuit. A pupil detection circuit detects images of pupils of the subject, based on the image of the face detected by the face detection circuit. A catch-light composition circuit composes different catch-lights on the images of the pupils detected by the pupil detection circuit, according to any one of the direction and expression of the face detected by the face expression detection circuit.

Abstract: A method of classifying an input image includes the initial steps of labeling an input image in accordance with a class and extracting at least one connected component from the input image. The method also includes the steps of calculating at least one feature of the input image and generating a model based on the at least one calculated feature. The method also includes the steps of repeating at least one of the previous steps for at least one other input image and comparing the at least one other input image with the model. The at least one other input image is classified in accordance with the class of the model if the at least one calculated feature of the at least one other input image is substantially similar to that of the model.

Abstract: An input-handwriting automatic transformation system capable of automatically transforming handwriting input to a font most similar to the input handwriting, the system including a recognizing unit recognizing handwriting input via an input pad; an extracting unit extracting a font most similar to the input handwriting from fonts stored in a memory; and a transforming unit comparing the font extracted by the extracting unit and the input handwriting and automatically transforming the extracted font to be most similar to the input handwriting.

Abstract: Locations of word images corresponding to words in a document image are ascertained. The word images are grouped into clusters. For each of multiple of the clusters, a respective compressed word image cluster is determined based on a joint compression of respective ones of the word images that are grouped into the cluster. The positions of the word images in the document image are associated with the respective ones of the compressed word image clusters corresponding to the clusters respectively containing the word images.

Abstract: A cell phone having distributed artificial intelligence services is provided. The cell phone includes a neural network for performing a first pass of object recognition on an image to identify objects of interest therein based on one or more criterion. The cell phone also includes a patch generator for deriving patches from the objects of interest. Each of the patches includes a portion of a respective one of the objects of interest. The cell phone additionally includes a transmitter for transmitting the patches to a server for further processing in place of an entirety of the image to reduce network traffic.

Abstract: A method for compressing an elevation map by means of digital image processing, according to which local and global parameters of an elevation map are calculated for sorting the elevation map into a corresponding category and the elevation map is categorized according to the calculated local and global parameters. Optimal parameters for compressing the elevation map are determined according to the calculated local and global parameters and according to the corresponding category and then the elevation map is compressed by applying on it digital image processing by using the determined optimal compression parameters.

Abstract: A character area is detected in a reduced layout and a character is detected in the character area, and a character size is compared to a predetermined size readable by a user when characters are printed. A detected character size determined to be smaller than the predetermined size is enlarged. An object adjacent to the detected character in an original document is detected and a content of the objects is determined. A layout position in which a relationship between the enlarged detected characters and the objects is kept is determined according to the determined content and the enlarged characters are arranged. The whole document can be output automatically in an easily viewable state while keeping a meaningful layout relationship between the objects.

Abstract: The pixels located at equal intervals in a shot image are converted into intensity components and applied with a differential filter. The number of the obtained variations that have values larger than or equal to a certain value among the obtained variations is counted. Then discrimination of a vehicle shot in the target image is executed on the basis of the counted value. The counted value varies depending on a size of a shot object in a shot image. The counted value is large when an occupied area of the object is large, and the counted value is small when the occupied area of the object is small. By so doing, it is advantageously possible to properly discriminate a vehicle.

Abstract: According to one embodiment, a postal indicium detection method includes detecting whether or not one of a first face and a second face of a sheet corresponds to a picture card specific face, detecting a postal indicium candidate from the second face based on a first detection result indicating that the first face corresponds to the picture card specific face, and detecting the postal indicium candidate from the first face based on a second detection result indicating that the second face corresponds to the picture card specific face.

Abstract: The invention relates to a method for comparing a first computer-aided 3D model with a second computer-aided 3D model, the first model being described by a first topology comprising vertices, edges and faces of the first model and the second model being described by a second topology comprising vertices, edges and faces of the second model, the method comprising the following steps: scanning the first topology and the second topology; deriving from the first topology a first topological map relating vertices, edges and faces of the first model and from the second topology a second topological map relating vertices, edges and faces of the second model; determining by means of the topological maps all possible mappings between vertices of the first model and vertices of the second model; using those mapped vertices to identify all possible mappings between edges of the first and edges of the second model; using those mapped edges to identify all possible mappings between faces of the first and faces of the seco

Abstract: A method is provided for decompressing an image originally compressed using a known compression technique. An original compressed image file is defined by a plurality of locations that correspond to image data. A set of smooth test images is created by iteratively applying a plurality of different size PIXON® kernels to each location of a starting image beginning with a smallest size kernel. Each smooth test image is compressed using the known compression technique, then compared against the original compressed image at each location of the image to determine a goodness-of-fit of the compressed smooth test image to each location of the original compressed image file. For each location, the smooth test image that has the largest size kernel that passes a pre-determined goodness-of-fit criterion is selected for a given location of the decompressed image. A compression method utilizes the PIXON® method for optimization of knot points within an image.

Abstract: In an electronic document of drawing descriptions of a page image and a character, it is desired that although a font data necessary for drawing the character is held in the electronic document, the size of the electronic document is minimized. Furthermore, it is desired to ensure visibility at the time of highlighting of search. There is generated an electronic document in which a document image, a plurality of character codes obtained by executing a character recognition processing with respect to the document image, and a plurality of kinds of glyph data to be utilized in common with respect to the plurality of character codes when drawing characters corresponding to the plurality of character codes are stored. The plurality of kinds of glyph data are selectively used when characters corresponding to the character codes are drawn. It is desirable that the glyph data be the one in a simple form.

Abstract: A method of recognizing an event depicted in an image from the image and a location information associated with the image is disclosed. The method includes acquiring the image and its associated location information; using the location information to acquire an aerial image(s) correlated to the location information; identifying the event using both the image and the acquired aerial image(s); and storing the event in association with the image for subsequent use.