Abstract: A method and system for searching multimedia contents are described. Embodiments of the invention utilize a multimedia user graphical interface (MGUI) and apply techniques of the Levenshtein algorithm for matching or searching one-dimensional strings to comparing graphical contents of 2D images.

Abstract: A system and method of machine learning that uses an inverse matrix of a reference similarity matrix as a transformation matrix. The reference similarity matrix relates a reference set of objects to themselves using a distance metric such as an image edit distance. The transformation matrix is used to improve the performance of query vectors in classifying or identifying digital representations of an unknown object. The query vector is a measure of similarity between the unknown object and the members of the reference set. Multiplying the query vector by the transformation matrix produces an improved query vector having improved similarity scores. The highest improved similarity score indicates the best match member of the reference set If the similarity score is high enough, the unknown object may either be classified as belonging to the same class, or recognized as being the same object, as the best match object.

Abstract: A system and method for rapidly searching large databases. A database is transformed into a similarity matrix using a similarity metric, such as an edit distance. A query object is compared to one member of the database using the same similarity metric, resulting in a similarity score. The row of the similarity matrix corresponding to the selected member is examined to find a best match similarity score. If the best match relates the selected member to itself, then the query object is identified as being the selected member, as long as it is above a threshold. If, not, the process is repeated using the other member of the database referred to by the best match. The process is repeated until the process converges, i.e. until the best match to the similarity score of the query object and the reference object is the element relating the reference object to itself.

Abstract: A system and method for comparing images by calculating an edit distance between the images using the results of matching portions of one image to the other. The first image is divided into blocks of pixels. For every block of pixels, a closest match is found in the second image. Substitution is equated to the sum of the errors in the closet matches. Deletion is equated to the percentage of the second image not involved in any of the matching. Insertion is equated to the percentage of the second image simultaneously involved in matching more than one block. The image edit distance may then be calculated as the weighted sum of the insertion, deletion and substitution. Recognition may be done by finding a minimum edit distance between an image of an unknown object and a set of reference images.

Abstract: A system and method for incorporating relevance feedback into an image search. A multimedia database is first searched using a domain specific similarity matrix (DSM), resulting in a list of image results ranked by their similarity to a query key-image. The results are then reexamined using a personalized similarity matrix (PSM). The PSM includes image that have been selected by the user, either in previous searches or from lists of candidate images. Similarity scores are calculated between the domain relevant images in the PSM and the results images. The new similarity scores are combined with the original similarity scores and the images is re-ranked into a final results list. Any images selected by the user may then be used to update and/or expand the PSM.

Abstract: A system and method for searching multimedia databases using a pictorial language, input via an iconic interface and making use of trained ontologies, i.e., trained data models. An iconic graphic user interface (GUI) allows a user to specify a pictorial query that may include one or more one or more key-images and optional text input. Similarities between the query key-images and images in a multimedia database based on a pictorial edit distance are used to select the images that are the closest match to the query. The result images are returned to the user in order of their relevance to the query.

Abstract: A system and method of searching multimedia databases using key images that returns image content ranked by degree of similarity to the key-images. A user accesses the search engine via a graphic user interface (GUI) that allows the user to enter key-images using drag-and-drop technology. The user may also enhance the description of the key-images using text based input, or used text based input to call up exemplary images. The user may also use Boolean logic to combine the key-images, and the key-images and text input, into search strings. The search strings may then be used to find matches from pre-indexed directories that have been created by analyzing content in advance.

Abstract: A method and system for tracking one or more targets are described. The method includes the step of selecting a first template having a first image of a target and cyclically repeated steps of accumulating new images of the target, producing updated templates containing the new images, and tracking the target using the updated templates. Embodiments of the method use techniques directed to detection and mitigation of target occlusion events.

Abstract: The present invention relates to mitigating the effects of distortions in pulse shape and energy of spectrum from gamma ray and/or neutron detectors. In one embodiment, the present invention uses pattern recognition techniques to identify and separate various distortions in the pulse shape and energy to obtain a better estimate of the true number of pulses of specific energies which are characteristic of the radioactive materials present. Additionally, autoregressive models, pulse filtering and pattern recognition methods are used to obtain more accurate pulse characterization. In one embodiment, the present invention relates to a method of detecting a material comprising the steps of deriving an energy spectrum from nuclear radiation detected from the material, processing the energy spectrum for enhancing the energy spectrum, extracting one or more features from the enhanced energy spectrum, and classifying the material based on the one or more features.

Abstract: An encoding method for reducing motion video data sizes using a multitude of variable-sized data blocks that are derived from spatial translation vectors and the motion field of an image. Using variable-block sizes to characterize the groups of picture elements (pixels) in an image frame allows for the inter-mixing of coarse (for static areas) and fine (for areas of complex motion) resolution data descriptions in the same descriptive data block. A comparison of motion event areas is made between successive video frames, and a motion displacement vector is calculated for each pixel location in the frame. A data tree is constructed from these pixel motion vectors and is pruned to eliminate static areas. The remaining leaves of the pruned tree are encoded differentially and applied to a lossless arithmetic encoder to provide a significantly reduced data block that still retains the highest resolution of the image.