Reconstitution of images

• US 4,649,568 A
• Filed: 10/22/1984
• Issued: 03/10/1987
• Est. Priority Date: 10/22/1984
• Status: Expired due to Fees

• Current Assignee: Polaroid Corporation (PLR IP Holdings LLC)
• Original Assignee: Polaroid Corporation (PLR IP Holdings LLC)
• Inventors: Kiesel, Kenneth C., Wray, William R.
• Primary Examiner(s): Ciarlante, Anthony V.
• Application Number:
  US06/663,719
• Time in Patent Office: 869 Days
• Field of Search: 382/41, 382/47, 382/44, 358/137, 358/140
• US Class Current: 382/162
• CPC Class Codes: G06T 3/4007 based on interpolation, e.g. bilinear interpolation image demosaicing G06T3/4015; edge-driven or edge-based scaling G06T3/403
• CPC Class Codes: G06T 5/94 based on local image properties, e.g. for local contrast enhancement

Abstract

Improvements in a full resolution image A are effected with an economical use of computer processing equipment and time by determining a field of improvement functions derived from processing a coarse representation of the original full resolution image. The coarse representation B is subjected to lightness field analysis to obtain an improved coarse image C, from which a coarse improvement field D is extracted by subtracting the original coarse representation. To obtain a full resolution improvement field E, the coarse improvement function is smoothly expanded by interpolation to the dimensions of the full resolution image. By modifying the original full resolution image field A with the values of the full resolution improvement field E, an improved full resolution image is obtained.

Claims

1. An image improvement method comprising the steps of

determining a radiance field representative of a comparatively fine resolution image-wise distribution of radiation from an original scene, and

determining an improved image field by modifying said radiance field with an improvement function derived from the processing of a regular array of light values of a comparatively coarse resolution representation of said original scene.

2. The method of claim 1 wherein

said method is repeated for each one of a plurality of ranges of wavelengths of radiation from the original scene to derive final image fields defining an improved multicolor image of said original scene.

3. An image analysis and reconstitution method which comprises

developing a coarse resolution improvement function from the processing of relatively coarse resolution radiance field representative of an image-wise distribution of radiance from an original scene, and

modifying a comparatively fine resolution radiance field representative of an image-wise distribution of radiance from said scene with a full resolution improvement function derived by smoothly expanding said coarse resolution improvement function to dimensions equivalent to those of said comparatively fine resolution radiance field.

4. An image improvement method comprising:

generating a finely divided and a coarsely divided radiance field, each representing an image-wise distribution of radiation within a selected range of wavelengths from an original scene;

processing said coarsely divided radiance field according to a predetermined analytical program to derive a coarse improvement function field;

smoothing said coarse improvement function field by interpolation to develop a smooth improvement function field equivalent to the dimensions of said finely divided radiance field; and

modifying each element of said finely divided radiance field by the corresponding element of said smooth improvement function field to determine an improved finely divided image of said original scene.

5. The method of claim 4 wherein the processing of the coarsely divided radiance field comprises:

the selected comparison of each element thereof with other elements of said coarsely divided radiance field to obtain a processed coarse image field having new comparative values for each of such elements responsive to the radiance values from substantially al other portions of said coarsely divided radiance field; and

wherein said coarsely divided radiance field is subtracted from said processed coarse image field to obtain the coarse improvement function field.

6. An image analysis and reconstitution method comprising:

determining a finely-divided radiance field representative of an image-wise distribution of radiation from an original scene,

generating a coarse radiance field representative of said original scene containing significantly fewer image elements than those of said finely divided radiance field,

processing said coarse radiance field according to a defined program to derive a processed coarse field,

comparing said coarse radiance field with said processed coarse field to determine a coarse improvement function field,

expanding and interpolating said coarse improvement function field to dimensions equivalent to the number of image elements in said finely divided radiance field to obtain a full resolution improvement function, and

modifying each of the elements of said finely divided radiance field by the corresponding element of said full resolution improvement function to determine an improved final image field.

7. An image analysis and reconstitution method comprising:

generating a relatively coarse divided radiance field representative of an image-wise distribution of radiation from an original scene,

processing said coarsely divided radiance field to determine a coarse improvement function field,

generating a relatively finely divided radiance field representative of an image-wise distribution of radiation from said scene with substantially higher resolution than that of said coarsely divided radiance field,

smoothly interpolating said coarse improvement function to derive a finely divided improvement function substantially equivalent to the dimensions of said finely divided radiance field, and

modifying each element of said finely divided radiance field in accordance with the value of the corresponding element of said finely divided improvement function to obtain a final image field for use in generating an improved image of said original subject.

8. An image analysis and reconstitution method comprising the steps of

developing a coarse resolution improvement function from the processing of a relatively coarse resolution radiance field representative of an image-wise distribution of radiance from an original scene,

smoothly expanding said coarse resolution improvement function to a full resolution improvement function defined on dimensions equivalent to those of a comparatively fine resolution radiance field representative of an image-wise distribution of radiance from said scene, and

modifying a comparatively fine resolution radiance field with said full resolution improvement function.

9. A method for improving a first radiance field having a radiance value defined at each point of a high resolution representation of a scene, wherein the method comprises the steps of

sensing, with a coarse resolution light sensing array, a radiance field representative of the scene so as to develop a coarse resolution radiance field having a single value on each pixel of the array,

computing a coarse radiance field improvement function as a single-valued function on each pixel of the array,

smoothing said coarse radiance field improvement function to obtain a smoothed improvement function that is smoothly defined on the points of the high resolution representation, and

modifying the first radiance field by said smooth improvement function to obtain an improved representation of the scene.

10. The method of claim 9 including the further step of controlling with a computer the distribution of light on a screen for displaying said improved representation of the scene.