Provided is a data management system which includes a data acquisition unit configured to acquire measurement data obtained by measuring a measurement target, a data storage unit configured to store the acquired measurement data, a data size reduction unit configured to deletes at least a part of the stored measurement data to reduce a data size of the measurement data, and a data compression unit configured to perform data compression on the measurement data reduced in data size.

One example method includes file specific compression selection. Compression metrics are generated for a chunk of a file. Using a set of training data, the compression metrics are corrected using a correction factor to determine estimated file compression metrics. A compressor is then selected to compress the file based on at least the estimated file compression metrics.

A memory includes, in one embodiment, one or more storage elements; read/write circuitry; and compressed bit circuitry. The read/write circuitry is configured to read a set of hard bits from the one or more storage elements, and sense a set of soft bits while reading the set of hard bits from the one or more storage elements, the set of soft bits having a first fixed size, and the set of soft bits indicating a reliability of the set of hard bits. The compressed soft bit circuitry is configured to generate, with a fixed size soft bit lossy compression algorithm, a fixed size compressed soft bits by compressing the set of soft bits, the fixed size compressed soft bits having a second fixed size that is smaller than the first fixed size, and output the fixed size compressed soft bits to a memory-to-controller bus.

Methods and apparatus for a sensing system having a focal plane array having an n×m array of sensing elements and a single output pixel and a mask to select or deselect ones of the sensing elements in the array to form patterns, wherein the mask forms a part of the focal plane array.

A memory includes, in one embodiment, one or more storage elements; read/write circuitry; and compressed bit circuitry. The read/write circuitry is configured to read a set of hard bits from the one or more storage elements, and sense a set of soft bits while reading the set of hard bits from the one or more storage elements, the set of soft bits having a first fixed size, and the set of soft bits indicating a reliability of the set of hard bits. The compressed soft bit circuitry is configured to generate, with a fixed size soft bit lossy compression algorithm, a fixed size compressed soft bits by compressing the set of soft bits, the fixed size compressed soft bits having a second fixed size that is smaller than the first fixed size, and output the fixed size compressed soft bits to a memory-to-controller bus.

There is provided a calibration device including: a calibration signal supply unit configured to supply, as a calibration input signal, a multitone signal having tones in a plurality of frequency bands to a converter configured to multiply an input signal by each of a plurality of signal patterns and limit a band to obtain each of a plurality of bandpass signals, and reconstruct an output signal in accordance with an input signal from the plurality of bandpass signals; a calibration bandpass signal acquisition unit configured to acquire a plurality of calibration bandpass signals obtained by the converter in response to the multitone signal; and a calibration processing unit configured to calibrate a parameter for the reconstruction in the converter based on the plurality of calibration bandpass signals.

Methods and apparatus for a sensing system having a focal plane array having an n×m array of sensing elements and a single output pixel and a mask to select or deselect ones of the sensing elements in the array to form patterns, wherein the mask forms a part of the focal plane array.

Techniques for classifier generalization in a supervised learning process using input encoding are provided. In one aspect, a method for classification generalization includes: encoding original input features from at least one input sample {right arrow over (x)}.sub.S with a uniquely decodable code using an encoder E(⋅) to produce encoded input features E({right arrow over (x)}.sub.S), wherein the at least one input sample {right arrow over (x)}.sub.S comprises uncoded input features; feeding the uncoded input features and the encoded input features E({right arrow over (x)}.sub.S) to a base model to build an encoded model; and learning a classification function {tilde over (C)}.sub.E(⋅) using the encoded model, wherein the classification function {tilde over (C)}.sub.E(⋅) learned using the encoded model is more general than that learned using the uncoded input features alone.

The present invention relates to a compressed sensing apparatus for compressed sensing of a set function consisting of a plurality of input sets containing a group of data. The apparatus includes: a plurality of sensing units acquiring a group of sampling data representing a plurality of sampling sets selected out of the plurality of input sets; a compression and computation unit enabling a compression to the group of data based on the group of sampling data in accordance with a Fourier basis set generated on the basis of the plurality of input sets and sampling sets, and a computation to compute a Fourier coefficient set based on a sparse regression technique which is in relation with the Fourier basis set; and a reconstruction unit predicting the group of data based on the Fourier coefficient set.

The present disclosure is directed to a snapshot multiframe imager having an aperture element having at least one aperture, an adjacently positioned random mask, an imaging element and a computer. The random mask has a plurality of micron scale apertures and receives light passing through the aperture element, which represents the spatial information from the scene being imaged, and generates a plurality of image frames encoded in a spatial domain. The imaging element may operate in a drift-scan mode receives the encoded image frames and generates a streaked pattern of electrons representing a plurality of images of the scene at a plurality of different times. The computer analyzes the streaked pattern of electrons and mathematically reconstructs the plurality of images.