Accelerated signal encoding and reconstruction using pixon method
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G06K-009/56
G06K-009/40
출원번호
US-0308450
(2002-12-02)
발명자
/ 주소
Puetter, Richard
Yahil, Amos
Pi?a, Robert
출원인 / 주소
The Regents of the University of California
대리인 / 주소
Procopio, Cory, Hargreaves &
인용정보
피인용 횟수 :
11인용 특허 :
9
초록▼
The method identifies a Pixon element, which is a fundamental and indivisible unit of information, and a Pixon basis, which is the set of possible functions from which the Pixon elements are selected. The actual Pixon elements selected from this basis during the reconstruction process represents the
The method identifies a Pixon element, which is a fundamental and indivisible unit of information, and a Pixon basis, which is the set of possible functions from which the Pixon elements are selected. The actual Pixon elements selected from this basis during the reconstruction process represents the smallest number of such units required to fit the data and representing the minimum number of parameters necessary to specify the image. The Pixon kernels can have arbitrary properties (e.g., shape, size, and/or position) as needed to best fit the data.
대표청구항▼
1. A method for extracting information from a data signal having a plurality of data locations, the method comprising:(a) defining a plurality of different basis function sets, each different basis function set comprising a plurality of identical basis functions, wherein each basis function within a
1. A method for extracting information from a data signal having a plurality of data locations, the method comprising:(a) defining a plurality of different basis function sets, each different basis function set comprising a plurality of identical basis functions, wherein each basis function within a basis function set encompasses a portion of the plurality of data locations so that the plurality of basis functions of a basis function set encompasses the entire data signal, and wherein the basis functions within each different basis function set encompass a different quantity of data locations within the data signal compared to the basis functions within other different basis function sets; (b) decomposing the data signal with the plurality of different basis function sets; (c) determining a signal-to-noise ratio and a goodness-of-fit for each basis function by mapping the function into data space and comparing the mapped function to the portion of the plurality of data locations within the basis function; (d) comparing the signal-to-noise ratio and the goodness-of-fit for each basis function against predetermined pass criteria; (e) selecting at each data location an optimal basis function from the plurality of different basis function sets wherein the optimal basis function is the broadest basis function that meets the predetermined pass criteria at that data location; and (f) generating a basis function map comprising a plurality of optimal basis functions, wherein each optimal basis function encompasses at least one data location within the data signal so that the entire data signal is encompassed by the plurality of optimal basis functions. 2. The method of claim 1, wherein the data signal comprises a detected image and the information to be extracted comprises an image having minimized degrees of freedom relative to the detected image.3. The method of claim 2, wherein the detected image is arranged in an image grid corresponding to a plurality of pixels in an image detector and the smallest basis function corresponds to one pixel in the image detector.4. The method of claim 3, wherein each basis function is centered on a pixel in the image detector.5. The method of claim 2, wherein each basis function has a shape selected from the group consisting of circles, ellipses, squares, polygons, parabolic, and Gaussian.6. The method of claim 1, wherein the data signal comprises time series data and the information to be extracted comprises an image having minimized degrees of freedom relative to the time series data.7. The method of claim 6, wherein the time series data comprises input from an array of infrared detectors.8. The method of claim 1, wherein the data signal comprises a signal which is to be modulated for RF transmission and the information to be extracted comprises a minimum number of parameters capable of describing the signal.9. The method of claim 1, wherein the data signal comprises a signal which is to be encoded for compression.10. The method of claim 1, wherein steps (b) and (c) are repeated hierarchically for each different basis function set beginning with the basis function set encompassing the largest portion of the plurality of data locations within the data signal and proceeding toward the basis function set encompassing the smallest portion of the plurality of data locations.11. The method of claim 10, wherein repetition of steps (b) and (c) is terminated at the data location within the data signal once the optimal basis function corresponding to that data location is selected.12. The method of claim 1, wherein the different basis function sets cover a range spanning three to five octaves.13. The method of claim 1, wherein each basis function is radially symmetric.14. A method for extracting information from a data signal comprising:assembling a map comprising a plurality of optimal basis functions for modeling the information represented within the data signal, wherein the data signal has a plurality of data locations, and wherein the plurality of optimal basis functions is selected from a plurality of different basis function sets, each different basis function set encompassing a different quantity of data locations compared to other different basis functions of the plurality of different basis function sets, wherein each optimal basis function is selected by determining a local goodness-of-fit and a local signal-to-noise ratio for the plurality of different basis functions relative to the data locations encompassed by the basis function and, for each portion of the data signal, selecting as the optimal basis function the basis function that encompasses the greatest quantity of data locations and exceeds predetermined pass criteria. 15. The method of claim 14, wherein the predetermined pass criteria comprise signal-to-noise ratio and goodness-of-fit.16. The method of claim 14, wherein each different basis function has a shape selected from the group consisting of circles, ellipses, squares, polygons, parabolic, and Gaussian.17. The method of claim 14, wherein each different basis function is radially symmetric.18. The method of claim 14, wherein each different basis function set has a size that is logarithmically spaced from the basis function sizes of other basis function sets.19. The method of claim 18, wherein the basis function sizes for the different basis function sets cover a range spanning from three to five octaves.20. The method of claim 14, wherein the optimal basis function is selected by, until the predetermined criteria are met, repeatedly determining the local goodness-of-fit and the local signal-to-noise ratio for the plurality of different basis function sets relative to the data locations within the data signal that are encompassed by the basis function for each of the plurality of different basis function sets beginning with the basis function encompassing the largest quantity of data locations and proceeding toward the basis function encompassing the smallest quantity of data locations.21. The method of claim 14, wherein the data signal comprises a detected image and the information to be extracted comprises an image having minimized degrees of freedom relative to the detected image.22. The method of claim 21, wherein the detected image is arranged in an image grid corresponding to a plurality of pixels in an image detector and the basis function which encompasses the smallest quantity of data locations corresponds to one pixel in the image detector.23. The method of claim 22, wherein each basis function is centered on a pixel in the image detector.24. The method of claim 21, wherein each basis function has a shape selected from the group consisting of circles, ellipses, squares, polygons, parabolic, and Gaussian.25. The method of claim 14, wherein the data signal comprises time series data and the information to be extracted comprises an image having minimized degrees of freedom relative to the time series data.26. The method of claim 25, wherein the time series data comprises input from an array of infrared detectors.27. The method of claim 14, wherein the data signal comprises a signal which is to be modulated for RF transmission and the information to be extracted comprises a minimum number of parameters capable of describing the signal.28. The method of claim 14, wherein the data signal comprises a signal which is to be encoded for compression.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (9)
Richard Puetter ; Amos Yahil, Accelerated signal encoding and reconstruction using pixon method.
Johnson Steven A. (1916 E. Browning Ave. Salt Lake City UT 84108) Wiskin James W. (1021 Medical Towers Salt Lake City UT 84112) Borup David T. (1133 E. Sunnyside Ave. Salt Lake City UT 84103) Christe, Apparatus and method for imaging with wavefields using inverse scattering techniques.
Floyd ; Jr. Carey E. (Chapel Hill NC) Baydush Alan H. (Durham NC), X-ray procedure for removing scattered radiation and enhancing signal-to-noise ratio (SNR).
※ AI-Helper는 부적절한 답변을 할 수 있습니다.