Abstract: A method for determining the likelihood that a pixel in a video image sequence represents true motion in the image. Light from a scene is captured to yield an original image having a sequence of frames. A prototype image is created from the original image as a geometrically accurate representation of the scene which is substantially free of contamination. A tilt variance-Gaussian mixture model and/or multi-variate Gaussian model, based upon optical turbulence statistics associated with the scene, is developed. Each pixel in the original image sequence of frames is evaluated against the prototype image using a probability density function, to yield a likelihood that the respective pixel represents true motion.

Abstract: Systems and methods for transfer-to-transfer training using an imbalanced training dataset include reconfiguring an imbalanced training data corpus to a plurality of distinct class-balanced mini-corpora of training data, wherein the reconfiguring includes: (i) partitioning the imbalanced training data corpus into a plurality of mini-corpora of training data samples in which each distinct mini-corpus of the plurality of mini-corpora includes an entirety of the training data samples within the second subset of training data samples; and (ii) allocating an equal number of the training data samples of the first subset into each of the plurality of mini-corpora of training data samples; and transfer-to-transfer learning-based training a subject machine learning algorithm to a trained machine learning model based on implementing the transfer-to-transfer learning-based training using the plurality of distinct class-balanced mini-corpora, wherein in use, the trained machine learning model predicts a presence or a non

Abstract: Systems and methods for transfer-to-transfer training using an imbalanced training dataset include reconfiguring an imbalanced training data corpus to a plurality of distinct class-balanced mini-corpora of training data, wherein the reconfiguring includes: (i) partitioning the imbalanced training data corpus into a plurality of mini-corpora of training data samples in which each distinct mini-corpus of the plurality of mini-corpora includes an entirety of the training data samples within the second subset of training data samples; and (ii) allocating an equal number of the training data samples of the first subset into each of the plurality of mini-corpora of training data samples; and transfer-to-transfer learning-based training a subject machine learning algorithm to a trained machine learning model based on implementing the transfer-to-transfer learning-based training using the plurality of distinct class-balanced mini-corpora, wherein in use, the trained machine learning model predicts a presence or a non

Abstract: A method for refining shape estimates of detected abnormalities in medical images, particularly for the detection of pulmonary lesions in CT imagery is described. A cue point is refined prior to an initially segmentation of the lesion. A radial gradient is computed for points on an initial segmented lesion. Portions of the lesion with radial gradients deviating beyond a threshold angle are removed from the lesion. Registering imagery from more than one CT exam uses a high intensity structure, essentially bone, to provide coarse and fine alignment of a set of two-dimensional images is also described. A MIP image is formed from the three-dimensional images. A second MIP image is then formed from imagery. The second MIP image is correlated with the first MIP image to determine a preferred registration location.