Abstract: The present invention disclosed an iterative analyzing method, which can detect the lesion in the image quickly. In brief, the iterative analysis method of the medical image disclosed by the present invention is roughly as follows: first, the original spectral image cube is expanded into a spectral image cube by a method of nonlinear dimensional-expansion, and then detecting the target's subpixel by the method of constrained energy minimization to produce an abundance image; the abundance image is fed back to the spectral image cube for create another spectral image cube by the nonlinear method. Furthermore, the abundance image is only used for detecting the subpixel of the target and does not include any spatial information, so, in order to obtain the spatial information of spectral image, it obtains the spatial information around the subpixel by using a blurring tool such as a Gaussian filter.

Abstract: The present invention disclosed an iterative analyzing method, which can detect the lesion in the image quickly. In brief, the iterative analysis method of the medical image disclosed by the present invention is roughly as follows: first, the original spectral image cube is expanded into a spectral image cube by a method of nonlinear dimensional-expansion, and then detecting the target's subpixel by the method of constrained energy minimization to produce an abundance image; the abundance image is fed back to the spectral image cube for create another spectral image cube by the nonlinear method. Furthermore, the abundance image is only used for detecting the subpixel of the target and does not include any spatial information, so, in order to obtain the spatial information of spectral image, it obtains the spatial information around the subpixel by using a blurring tool such as a Gaussian filter.

Abstract: Techniques are provided herein for receiving at a video processing device color video frames comprising grayscale components and color components. The grayscale components corresponding to each of the color video frames are extracted as each of the color video frames is received to obtain grayscale video frames. The grayscale video frames are transmitted at a first transmission rate. Color components for selected color video frames are periodically fused with selected grayscale video frames to obtain fused color video frames, and the fused color video frames are transmitted at a second transmission rate interspersed with the grayscale video frames transmitted at the first transmission rate.

Abstract: Provided herein are algorithms and processes to extract endmembers from hyperspectral image data in real time. A Simplex Growing Algorithm is effective to estimate a p number of endmembers to be generated, to select one or more initial endmembers as a simplex of k members and to add a k+1 endmember to the simplex that yields a maximum simplex volume until k=p, thereby extracting one or more endmembers from the data. Alternatively, N-FINDR algorithms form an initial simplex set of p endmembers obtained from the hyperspectral image data, find a maximum volume of one or more initial p endmembers therewithin, replace one or more of the p endmembers within the simplex with one or more of the found p endmembers of maximum volume, and refind a maximum volume of p endmember(s) and replace p endmember(s) until no increase in p endmember(s) volume is found.

Abstract: Techniques for automatically detecting meniscus include receiving pixels from a scanning device directed to a knee of a subject. Also received is a region of interest that includes pixels that correspond to at least a portion of a meniscus of the knee. Without human intervention, a meniscus extraction threshold is determined based on pixel intensities in the region of interest. A meniscus object portion of the scan data is further determined without human intervention based on the meniscus extraction threshold and a geometrical constraint. Other techniques for automatically detecting meniscal tears includes receiving a meniscus object portion of scan data from a scanning device and receiving threshold data that indicates a meniscus extraction threshold. Without human intervention, a propensity for meniscal tears is determined based on the threshold data and the meniscus object portion of the scan data.

Abstract: Techniques are provided herein for receiving at a video processing device color video frames comprising grayscale components and color components. The grayscale components corresponding to each of the color video frames are extracted as each of the color video frames is received to obtain grayscale video frames. The grayscale video frames are transmitted at a first transmission rate. Color components for selected color video frames are periodically fused with selected grayscale video frames to obtain fused color video frames, and the fused color video frames are transmitted at a second transmission rate interspersed with the grayscale video frames transmitted at the first transmission rate.

Abstract: Provided herein are algorithms and processes to extract endmembers from hyperspectral image data in real time. A Simplex Growing Algorithm is effective to estimate a p number of endmembers to be generated, to select one or more initial endmembers as a simplex of k members and to add a k+1 endmember to the simplex that yields a maximum simplex volume until k=p, thereby extracting one or more endmembers from the data. Alternatively, N-FINDR algorithms form an initial simplex set of p endmembers obtained from the hyperspectral image data, find a maximum volume of one or more initial p endmembers therewithin, replace one or more of the p endmembers within the simplex with one or more of the found p endmembers of maximum volume, and refind a maximum volume of p endmember(s) and replace p endmember(s) until no increase in p endmember(s) volume is found.

Abstract: Techniques for automatically detecting meniscus include receiving pixels from a scanning device directed to a knee of a subject. Also received is a region of interest that includes pixels that correspond to at least a portion of a meniscus of the knee. Without human intervention, a meniscus extraction threshold is determined based on pixel intensities in the region of interest. A meniscus object portion of the scan data is further determined without human intervention based on the meniscus extraction threshold and a geometrical constraint. Other techniques for automatically detecting meniscal tears includes receiving a meniscus object portion of scan data from a scanning device and receiving threshold data that indicates a meniscus extraction threshold. Without human intervention, a propensity for meniscal tears is determined based on the threshold data and the meniscus object portion of the scan data.

Abstract: A Field Programmable Gate Arrays (FPGA) design uses a Coordinate Rotation DIgital Computer (CORDIC) algorithm that can convert a Givens rotation of a vector to a set of shift-add operations. The CORDIC algorithm can be easily implemented in hardware architecture, therefore in FPGA. Since the computation of the inverse of the data correlation matrix involves a series of Givens rotations, the utility of the CORDIC algorithm allows a causal Constrained Energy Minimization (CEM) to perform real-time processing in FPGA. An FPGA implementation of the causal CEM is described and its detailed architecture is also described.

Abstract: A Field Programmable Gate Arrays (FPGA) design uses a Coordinate Rotation DIgital Computer (CORDIC) algorithm that can convert a Givens rotation of a vector to a set of shift-add operations. The CORDIC algorithm can be easily implemented in hardware architecture, therefore in FPGA. Since the computation of the inverse of the data correlation matrix involves a series of Givens rotations, the utility of the CORDIC algorithm allows a causal Constrained Energy Minimization (CEM) to perform real-time processing in FPGA. An FPGA implementation of the causal CEM is described and its detailed architecture is also described.

Abstract: A method using CC and MLO views localized in a display in 3-D as a breast virtual model, incorporating, feature capturing using gradient, energy and entropy codes, calibration classification according to features using binary decision trees, nipple detection, and 3-D spatial coordinate transformation and display of breast clustered microcalcifications.

Abstract: The present invention relates to a new and useful Automated Pattern Recognition Device comprising a neural-network system, implemented on a general purpose computer, and capable of recognizing not only printed characters but also handwritten characters and other patterns in n-dimensions. The system incorporates novel feature extraction which generates an additional dimension from an n-dimensional input pattern, for example, a three-dimensional feature pattern from a two dimensional input pattern, resulting in shift-invariance, scale-invariance, and invariance to slight rotation.