Method and Apparatus for Signal Processing

• US 20130241633A1
• Filed: 09/12/2012
• Published: 09/19/2013
• Est. Priority Date: 09/13/2011
• Status: Active Grant

• Current Assignee: JEOL Limited
• Original Assignee: JEOL Limited
• Inventors: Honda, Kazuhiro
• Granted Patent: US 8,885,968 B2
• Time in Patent Office: Days
• US Class Current: 327/551
• CPC Class Codes: G06T 2207/10061 from scanning electron microscope
• CPC Class Codes: G06T 2207/30168 Image quality inspection
• CPC Class Codes: G06T 5/00 Image enhancement or restoration
• CPC Class Codes: H03K 5/1252 Suppression or limitation of noise or interference specially adapted for transmission systems H04B15/00, H04L25/08

Abstract

Signal processing method and apparatus having a first filter storage portion in which first filters are correlatively stored; a second filter storage portion in which second filters are correlatively stored; a first filter selection portion for selecting a first filter based on the power spectrum of the input image; a second filter selection portion for selecting a second filter based on the S/N of the input image; a third filter creation portion for creating a third filter by summing up the first and second filters; and a convolutional processing portion for convolving the input image using the created third filter.

Claims

1. A method for signal processing, comprising the steps of:

calculating a power spectrum of an input signal;

selecting a power spectrum approximate to the calculated power spectrum from a plurality of power spectra stored in a first filter storage portion;

selecting a first filter corresponding to the selected power spectrum from a plurality of first filters stored in the first filter storage portion;

calculating an S/N of the input signal;

selecting an S/N approximate to the calculated S/N from a plurality of S/Ns stored in a second filter storage portion;

selecting a second filter corresponding to the selected S/N from a plurality of second filters stored in the second filter storage portion;

creating a third filter by adding up the selected first and second filters; and

convolving the input signal using the created third filter.

2. A method for signal processing as set forth in claim 1,

wherein the first filters stored in said first filter storage portion have been created by a step of creating filters based on sample input signals and are correlated with power spectra of the sample input signals,

wherein the second filters stored in said second filter storage portion have been created by said step of creating filters and are correlated with S/Ns of the sample input signals, and

wherein during said step of creating filters, a filter of an arbitrary distribution function G₀ is expanded into a series of n filter elements (where n is an arbitrary number) each of which is the difference between two distribution functions, coefficients of the filter elements are set at will, the filter elements are multiplied by the coefficients, respectively, and summed up to create the first filters being a function of n (where n is the number of the filter elements) and of the coefficients of the filter elements, the distribution function G_n of the nth order obtained when expanding the filter into the n filter elements and the distribution function G₀ are combined together with arbitrary weight coefficients to create the second filters, and the value of n, the coefficients of the filter elements, and the weight coefficients are set based on the sample input signals.

3. A method for signal processing as set forth in claim 2, wherein during said step of creating filters, the value of number n, the coefficients of the filter elements, and the weight coefficients are set to maximize the S/Ns of the signals obtained by convolving the sample input signals using the created first and second filters.

4. A method for signal processing as set forth in any one of claims 2 and 3, wherein each of said distribution functions is a Gaussian function.

5. An apparatus for signal processing, comprising:

a first filter storage portion in which first filters created based on sample input signals and power spectra of the sample input signals are correlatively stored;

a second filter storage portion in which second filters created based on sample input signals and S/Ns of the sample input signals are correlatively stored;

a first filter selection portion for calculating a power spectrum of an input signal, selecting a power spectrum approximate to the calculated power spectrum from the power spectra stored in the first filter storage portion, and selecting a first filter corresponding to the selected power spectrum from the first filters stored in the first filter storage portion;

a second filter selection portion for calculating an S/N of the input signal, selecting an S/N approximate to the calculated S/N from the S/Ns stored in the second filter storage portion, and selecting a second filter corresponding to the selected S/N from the second filters stored in the second filter storage portion;

a third filter creation portion for creating a third filter by summing up the selected first and second filters; and

a convolutional processing portion for convolving the input signal using the created third filter.

6. An apparatus for signal processing as set forth in claim 5,

wherein there is further provided a filter creation portion for creating the first and second filters based on the sample input signals and calculating power spectra and S/Ns of the sample input signals, and

wherein said filter creation portion expands a filter of an arbitrary distribution function G₀ into a series of n filter elements (where n is an arbitrary number) each of which is the difference between two distribution functions, sets coefficients of the filter elements at will, multiples the filter elements by the coefficients, respectively, and sums up the multiplied filter elements to create the first filter that is a function of the number of the filter elements n and of the coefficients of the filter elements, combines together a distribution function G_n of the nth order obtained when expanding the filter into the n filter elements and the distribution function G₀ with an arbitrary weight coefficient to create the second filter, and sets the value of n, the coefficients of the filter elements, and the weight coefficient based on the sample input signals.

7. An apparatus for signal processing as set forth in claim 6, wherein said filter creation portion sets the value of the number n, the coefficients of the filter elements, and the weight coefficients to maximize the S/N of the signals obtained by convolving the sample input signals using the created first and second filters.

8. An apparatus for signal processing as set forth in any one of claims 6 and 7, wherein each of said distribution functions is a Gaussian function.