Methods and apparatuses for detecting anomalies using transform based compressed sensing matrices
View Patent ↗A measurement vector of compressive measurements is received. The measurement vector may be derived by applying a sensing matrix to a source signal. The sensing matrix may be derived from a frequency domain transform. At least one first feature vector is generated from the measurement vector. The first feature vector is an estimate of a second feature vector. The second feature vector is a feature vector that corresponds to a translation of the source signal. An anomaly is detected to in the source signal based on the first feature vector.
1. A method for processing a source signal to detect anomalies, the method comprising:
receiving, by a processor, a measurement vector of compressive measurements, the measurement vector being derived by applying a sensing matrix to the source signal, the sensing matrix being derived from a frequency domain transform;
generating, by the processor, at least one first feature vector from the measurement vector, the first feature vector being an estimate of a second feature vector, the second feature vector corresponding to a translation of the source signal; and
detecting, by the processor, an anomaly in the source signal based on the first feature vector;
wherein the source signal is obtained from one or more image frames.
2. The method of claim 1 , wherein generating the at least one first feature vector is done without reconstructing the source signal.
3. The method of claim 1 , wherein the generating includes computing an estimate of a third feature vector, the third feature vector corresponding to the measurement vector.
4. The method of claim 1 , wherein the generating includes computing linear combinations of the at least one first feature vector.
5. The method of claim 1 , wherein more than one first feature vectors are generated, each of the first feature vectors having a corresponding translation parameter, the method further comprising:
comparing a group of elements of a set of the first feature vectors with other elements of the set of first feature vectors, and
determining which elements in the group of the set of the first feature vectors are similar to the other elements of the set of first feature vectors.
6. The method of claim 1 , wherein the source signal is multi-dimensional and the translation parameter includes translations in one or more dimensions.
7. The method of claim 6 , further comprising:
performing edge detection by generating a linear combination of the at least one first feature vector,
the performing corresponding to subtracting a linear combination of neighboring pixels from pixels in the source signal.
8. The method of claim 6 , wherein the detecting an anomaly includes detecting motion in the one or more image frames.
9. The method of claim 8 , wherein more than one first feature vectors are generated, each of the first feature vectors having a corresponding translation parameter, the method further comprising:
comparing a group of elements of a set of the first feature vectors with other elements of the set of first feature vectors, and
determining which elements in the group of the set of the first feature vectors are similar to the other elements of the set of first feature vectors.
10. The method of claim 9 , further comprising:
determining at least one of a speed of the detected motion, a direction of the detected motion, and a certainty of the detected motion,
the determining being based on at least one of (i) the translation parameters associated with the elements determined to be similar to each other, and (ii) a degree of similarity of the elements determined to be similar to each other.
11. The method of claim 8 , wherein the detected motion triggers an action if the detected motion meets at least one criterion.
12. A server for processing a source signal to detect anomalies, the server configured to:
receive a measurement vector of compressive measurements, the measurement vector being derived by applying a sensing matrix to a source signal, the sensing matrix being derived from a frequency domain transform;
generate at least one first feature vector from the measurement vector, the first feature vector being an estimate of a second feature vector, the second feature vector corresponding to a translation of the source signal; and
detect an anomaly in the source signal based on the first feature vector;
wherein the source signal is obtained from one or more image frames.
13. An image detection system for detecting anomalies in a source signal, the system comprising:
an image detection device for obtaining a measurement vector for processing the source signal, the image detection device configured to,
generate a set of compressive measurements by applying a shift-preserving sensing matrix to a source signal, the sensing matrix being derived from a frequency domain transform, and
generate a measurement vector based on the set of compressive measurements; and
a server for processing the source signal, the server configured to, receive the measurement vector;
generate at least one first feature vector from the measurement vector, the first feature vector being an estimate of a second feature vector, the second feature vector corresponding to a translation of the source signal; and
detect an anomaly in the source signal based on the first feature vector.