Compressive imaging using approximate message passing with denoising

Patent metadata

Field	Value
Publication number	US-9607362-B2
Application number	US-201514713788-A
Country	US
Kind code	B2
Filing date	May 15, 2015
Priority date	May 16, 2014
Publication date	Mar 28, 2017
Grant date	Mar 28, 2017

Abstract

Various examples of methods and systems are provided for compressive imaging using approximate message passing with denoising. According to an aspect, a method includes applying an approximate message passing (AMP) conversion framework to a plurality of substantially linear measurements for conversion into a plurality of scalar measurements. A denoiser algorithm can be applied to the plurality of scalar measurements to generate a plurality of denoised scalar measurements. Further, a conversion term can be applied to the plurality of denoised scalar measurements for converting the plurality of denoised scalar measurements to a plurality of denoised substantially linear measurements. The plurality of substantially linear measurements can represent two-dimensional or three-dimensional signals.

First claim

Therefore, at least the following is claimed:

1. A method, comprising:
applying, using processing circuitry including a processor and memory, an approximate message passing (AMP) conversion framework to a plurality of substantially linear measurements of a multi-dimensional signal for conversion into a plurality of scalar measurements, where the multi-dimensional signal comprises a two-dimensional image or a three-dimensional hyper spectral image or volume;
applying, using the processing circuitry, a non-scalar denoiser algorithm to the plurality of scalar measurements to generate a plurality of denoised scalar measurements; and
applying, using the processing circuitry, a conversion term to the plurality of denoised scalar measurements for converting the plurality of denoised scalar measurements to a plurality of denoised substantially linear measurements.

2. The method-of claim 1 , wherein the non-scalar denoiser algorithm is a dictionary learning based image denoiser.

3. The method of claim 2 , wherein the dictionary learning based image denoiser is a sparsity based simultaneous denoising and interpolation (SBSDI) algorithm.

4. The method of claim 1 , wherein a two-dimensional signal is denoised using a block-matching and 3-D filtering (BM3D) algorithm.

5. The method of claim 1 , wherein the plurality of substantially linear measurements represents a three-dimensional signal and the applied non-scalar denoiser algorithm is a block matching 4-D (BM4D) filtering image denoiser algorithm.

6. The method of claim 1 , wherein the non-scalar denoiser algorithm uses an adaptive Wiener filter wavelet-based denoiser.

7. The method of claim 6 , further comprising damping to convert the plurality of substantially linear measurements.

8. The method of claim 1 , wherein the denoiser algorithm is optimized for p-norm errors.

9. A method, comprising:
applying, using processing circuitry including a processor and memory, an approximate message passing (AMP) conversion framework to a plurality of substantially linear measurements for conversion into a plurality of scalar measurements, wherein the plurality of substantially linear measurements represents a three-dimensional signal that corresponds to a hyperspectral image;
applying, using the processing circuitry, a denoiser algorithm to the plurality of scalar measurements to generate a plurality of denoised scalar measurements; and
applying, using the processing circuitry, a conversion term to the plurality of denoised scalar measurements for converting the plurality of denoised scalar measurements to a plurality of denoised substantially linear measurements.

10. The method of claim 9 , wherein the applied denoiser algorithm is a block matching 4-D (BM4D) filtering image denoiser algorithm.

11. A method, comprising:
applying, using processing circuitry including a processor and memory, an approximate message passing (AMP) conversion framework to a plurality of substantially linear measurements for conversion into a plurality of scalar measurements;
applying, using the processing circuitry, a denoiser algorithm to the plurality of scalar measurements to generate a plurality of denoised scalar measurements, wherein the denoiser algorithm is a three-dimensional denoiser algorithm that applies a two-dimensional wavelet-based and one-dimensional discrete cosine transform prior to denoising; and
applying, using the processing circuitry, a conversion term to the plurality of denoised scalar measurements for converting the plurality of denoised scalar measurements to a plurality of denoised substantially linear measurements.

12. The method of claim 11 , wherein the three-dimensional denoiser algorithm is an adaptive Wiener filter configured for wavelet coefficients having a non-zero mean.

13. The method of claim 12 , further comprising damping to convert the plurality of substantially linear measurements.

14. A method, comprising:
applying, using processing circuitry including a processor and memory, an approximate message passing (AMP) conversion framework to a plurality of substantially linear measurements for conversion into a plurality of scalar measurements;
applying, using the processing circuitry, a denoiser algorithm to the plurality of scalar measurements to generate a plurality of denoised scalar measurements; and
applying, using the processing circuitry, a conversion term to the plurality of denoised scalar measurements for converting the plurality of denoised scalar measurements to a plurality of denoised substantially linear measurements, wherein an Onsager correction term portion of the conversion term is approximated.

Assignees

• Univ North Carolina State

Inventors

• Baron Dror
• Tan Jin
• Ma Yanting

Classifications

• G06T9/00

  Image coding (bandwidth or redundancy reduction for static pictures H04N1/41; coding or decoding of static colour picture signals H04N1/64; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00) · CPC title

• H04N19/90

  using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals · CPC title

• H04N19/85

  using pre-processing or post-processing specially adapted for video compression · CPC title

• G06T2207/20064

  Wavelet transform [DWT] · CPC title

• H04N19/80

  Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation (H04N19/635, H04N19/86 take precedence) · CPC title