Coded image system and method thereof

Patent metadata

Field	Value
Publication number	US-9445115-B2
Application number	US-201414547485-A
Country	US
Kind code	B2
Filing date	Nov 19, 2014
Priority date	Nov 21, 2013
Publication date	Sep 13, 2016
Grant date	Sep 13, 2016

Abstract

A coded image system enables a user to recover an object scene from an encoded image. The coded the image system comprises: a physical object scene; an encoding logic creating an encoded image, the encoding logic including a programmable aperture mask spaced a distance away from the physical object scene to encode the encoded image via a doubly-Toeplitz matrix, the doubly-Toeplitz matrix including two one-dimensional vectors; and a decoding logic operatively coupled to the programmable aperture mask to decode the encoded image recovering a visual representation of the object scene from the encoded image.

First claim

What is claimed:

1. A coded image system recovering an object scene from an encoded image, the image system comprising:
(a) an encoding logic creating an encoded image, the encoding logic including a doubly-Toeplitz matrix produced by two one-dimensional vectors, and an electronic device operatively coupled with the matrix adapted to be spaced a distance away from a physical object scene; and
(b) a decoding logic operatively coupled to the encoding logic to decode the encoded image recovering a visual representation of the object scene from the encoded image;
wherein the encoding logic permits that partial encoding may be inversed;
wherein the decoding logic comprises:
a reconstruction logic operatively coupled to the encoding logic to digitally reconstruct the object scene having a byte-size larger than the encoded image,
wherein the reconstruction logic comprises:
an imager including a dynamic range to unfold edges of the encoded image from a center without significant loss of information.

2. The coded image system of claim 1 , wherein the doubly-Toeplitz matrix is realized in a medium selected from a group comprising a programmable aperture mask, a non-programmable aperture mask, a liquid crystal spatial light modulator, an eyelid array and a self-supporting printed member.

3. The coded image system of claim 1 , wherein the decoding logic comprises:
a computational time period to decode the encoded image, wherein the time period is less than one second for images that are one hundred megapixel in size when a 10 Gigaflops processor is employed.

4. The coded image system of claim 3 , wherein the time period is less than ten milliseconds for images that are one megapixel in size when a 10 Gigaflops processor is employed.

5. The coded image system of claim 1 , wherein the decoding logic comprises:
a filtering logic operatively coupled to the encoding logic to remove pixel-to-pixel response variations caused by one or more of the following: dark noise, fixed-pattern noise, and bad pixels.

6. The coded image system of claim 5 , wherein the filtering logic comprises:
averaging logic to average multiple frames of imagery to mitigate random noise obtained by the encoding logic.

7. The coded image system of claim 1 , wherein the decoding logic comprises:
a regularization logic operatively coupled to the encoding logic to regularize the doubly-Toeplitz encoded image.

8. The coded image system of claim 7 , wherein the doubly-Toeplitz matrix is on a programmable aperture mask, and wherein the regularization logic comprises a Tikhonov regularization method.

9. The coded image system of claim 1 , wherein the decoding logic comprises:
a regularization logic operatively coupled to the programmable encoding logic obtains a normal equation that reduces noise effects including at least one photon counting and read noise;
wherein the two one-dimensional vectors are non-square defining an ill-posed matrix; and
wherein the regularization logic corrects the ill-posed matrix so that it may be inversed.

10. The coded image system of claim 9 , wherein the regularization logic includes a normal algorithm to correct the ill posed matrix so that it may be inversed.

11. The coded image system of claim 1 , wherein the reconstruction logic further comprises:
a spatial frequency spectra of the doubly-Toeplitz matrix free of any zeros.

12. The coded image system of claim 1 , wherein the reconstruction logic further comprises:
an optimization logic to estimate parameters to obtain a scene reconstruction, the optimization logic including at least one Tikhonov regularization constant to reduce noise during the reconstruction of the object scene.

13. The coded image system of claim 1 , further comprising:
a monitor to display the visual representation of the recovered object scene and wherein the physical object scene is non-uniform.

14. The coded image system of claim 1 , further comprising:
a pattern within one mask length unit from an edge of the reconstructed image; and
wherein the decoding logic includes a normalizing logic to normalize the pattern out of the reconstructed image by removing brightness differences accrued near the edge of the reconstructed image.

15. The coded image system of claim 14 , wherein the normalizing logic further comprises:
a estimated object (O Est );
a white estimate (W Est );
a normalized object estimate (O Norm ) and
wherein dividing the estimated object (O Est ) by the white estimate (W Est ) element-by-element to obtains the normalized object estimate (O Norm ).

16. The coded image system of claim 1 , wherein the decoding logic comprises:
an iteration logic operatively coupled to the encoding logic to reduce numerical round-off errors in inversions of a matrix programmed in the aperture mask.

17. The coded image system of claim 16 , wherein the iteration logic includes an iteration stopping logic comprising one of the following: a priori criteria and a posteriori criteria.

18. The coded image system of claim 1 , further comprising:
one of an imager and a camera operatively coupled to programmable aperture mask of the encoding logic to capture visual input of the physical object scene therethrough.

19. The coded image system of claim 1 , wherein the doubly-Toeplitz matrix permits radiation through openings defined by the matrix and blocks radiation through non-openings defined by the matrix, wherein the radiation is selected from the group comprising visible light, x-rays, gamma rays, UV light, and infrared light.

20. A method comprising the steps of:
encoding an image through a doubly-Toeplitz matrix coupled to a camera to create an encoded image, wherein the encoding permits that partial encoding may be inversed;
decoding the encoded image to recover a visual representation of an object scene, wherein the decoding digitally reconstructs the object scene,

wherein the reconstructing comprises,
imaging using an imager including a dynamic range to unfold edges of the encoded image from a center without significant loss of information; and
displaying the recovered object scene in a monitor, wherein the recovered object scene is larger than the encoded image.

21. The method of claim 20 , wherein the step of decoding is accomplished via Tikhonov regularization within a computational time period to decode less than one second for images that are one hundred megapixel in size.

22. The method of claim 20 , wherein the step of decoding includes the step of:
filtering the doubly-Toeplitz encoded image to remove pixel-to-pixel response variations caused by one or more of the following: dark noise, fixed-pattern noise, and bad pixels;
regularizing the doubly-Toeplitz encoded image via a Tikhonov regularization method; and
reconstructing the object scene as a decoded image larger than the encoded image.

23. The method of claim 22 , further comprising the steps of:
iterating the encoded image to reduce numerical round-off errors in inversions of the doubly-Toeplitz matrix in the aperture mask.

24. The method of claim 20 , wherein the doubly-Toeplitz matrix is realized in a medium selected from a group comprising a programmable aperture mask, a non-programmable aperture mask, a liquid crystal spatial light modulator, an eyelid array and a self-supporting printed member.

Assignees

• Bae Sys Inf & Elect Sys Integ

Inventors

• Deweert Michael J
• Farm Brian P
• Topping Miles

Classifications

• H04N19/44 Primary

  Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder · CPC title

• H04N19/42

  characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation (H04N19/635 takes precedence) · CPC title