Abstract: Systems and methods are disclosed for non-orthogonal encryption of data such that artificial neural networks are trained directly on the encrypted data and testing / inference is performed directly on encrypted data. For use with artificial neural networks, original input tensors are programmatically subjected to a sequence of encryption steps: a method for padding the feature set with randomly chosen sets of values; a method for perturbation of the data by altering elements of the feature set (the values stored in the dimensions of input vectors); a method that applies a fixed index shuffle of the data elements in the feature set or applies a random orthogonal matrix transformation of the data elements; and a method for applying convolutions (filters) and pooling (downsampling) to the fixed shuffled data such that informational structure is preserved.

Abstract: Systems and methods are disclosed for securely storing and programmatic searching of data files. The invention includes storing datafiles in fixed-sized packets of data that are subjected to fixed shuffle of the data elements (with or without prior fixed padding and with or without prior fixed perturbation), and target search elements are identified by way of the fixed shuffled indices, which are encrypted by conventional methods and decrypted for search by credentialed users.

Abstract: Systems and methods are disclosed for encrypting data such that artificial neural networks may be trained directly on the encrypted data and may be performed directly on the encrypted data. For use with artificial neural networks, original feature sets of input vectors are programmatically subjected to a sequence of encryption steps: a method for fixed padding of the feature set with a randomly chosen set of values that are fixed across the feature set (adding a fixed number of dimensions to each input vector); a method for fixed perturbation of the data by perturbing of each element in the feature set (the value stored in each dimension of an input vector) such that all examples of the feature set are perturbed in the same way; a method that applies a fixed shuffle of the data elements in the feature set; and a method for applying convolutions (filters) and pooling (downsampling) to the shuffled data such that informational structure is preserved.

Abstract: A technique manages medical image data. The technique involves receiving an original medical imaging study that includes a set of medical images and embedded protected patient information. The technique further involves performing a protected patient information removal operation which generates a cloud-storable medical imaging study from the original medical imaging study. The cloud-storable medical imaging study includes the set of medical images but omitting the embedded protected patient information. The technique further involves storing the cloud-storable medical imaging study in a cloud-based medical image repository among other cloud-storable medical image studies. The cloud-based medical image repository is external to the computerized device and public network accessible.

Abstract: The present invention teaches a method wherein medical imaging studies are transformed from Identifiable Imaging Studies into Cleared Imaging Studies that can be legally and securely shared, either by automated removal of protected health information (PHI) or by verification that the studies belong to a Patient or Patient's Legal Representative and that a valid effective waiver, such as a HIPAA Waiver, is on file. Cleared Imaging Studies are shared by cloud-based storage and transmission to one or more Third Parties using one or more network-enabled devices. Methods are also provided for one or more users of network-enabled devices to view medical imaging studies simultaneously, with caching of imaging studies in local devices and separation of additional data streams used for collaborative image viewing to ensure that medical images are not degraded.