Foveated image coding system and method for image bandwidth reduction
First Claim
1. A method for forming a compressed image, the method comprising:
- acquiring an image;
reducing the image with a Foveated Laplacian pyramid algorithm to form a plurality of down-sampled images, the down-sampled images decreasing in resolution from a highest to a lowest resolution;
designating a subset of each of the plurality of down-sampled images to form a plurality of subset images; and
processing the lowest resolution down-sampled image with the plurality of subset images to form the compressed image.
2 Assignments
0 Petitions
Accused Products
Abstract
A foveated imaging system, which can be implemented on a general purpose computer and greatly reduces the transmission bandwidth of images has been developed. This system has demonstrated that significant reductions in bandwidth can be achieved while still maintaining access to high detail at any point in an image. The system is implemented with conventional computer, display, and camera hardware. It utilizes novel algorithms for image coding and decoding that are superior both in degree of compression and in perceived image quality and is more flexible and adaptable to different bandwidth requirements and communications applications than previous systems. The system utilizes novel methods of incorporating human perceptual properties into the coding the decoding algorithms providing superior foveation. One version of the system includes a simple, inexpensive, parallel pipeline architecture, which enhances the capability for conventional and foveated data compression. Included are novel applications of foveated imaging in the transmission of pre-recorded video (without eye tracking), and in the use of alternate pointing devices for foveation.
-
Citations
13 Claims
-
1. A method for forming a compressed image, the method comprising:
-
acquiring an image;
reducing the image with a Foveated Laplacian pyramid algorithm to form a plurality of down-sampled images, the down-sampled images decreasing in resolution from a highest to a lowest resolution;
designating a subset of each of the plurality of down-sampled images to form a plurality of subset images; and
processing the lowest resolution down-sampled image with the plurality of subset images to form the compressed image. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
expanding the plurality of subset images to form a plurality of expanded subset images; and
subtracting the plurality of expanded subset images from the plurality of down-sampled images to form a plurality of difference images.
-
-
4. The method of claim 3, wherein the processing further comprises:
-
thresholding the plurality of difference images to form a plurality of threshold images;
quantizing the plurality of threshold images;
compressing the plurality of threshold images using a lossless compression algorithm to form a plurality of compression images;
compressing a down-sampled image using a lossless compression algorithm to form a compressed down-sampled image; and
transmitting the compressed down-sampled image with the plurality of compression images.
-
-
5. The method of claim 1, wherein the designating corresponds to one or more fixation points.
-
6. The method of claim 1, wherein the designating corresponds one or more screen coordinates of a pointing device.
-
7. The method of claim 6, wherein the pointing device comprises a mouse.
-
8. The method of claim 1, wherein the compressed image comprises a video image.
-
9. A method for forming a compressed image in real-time, the method comprising:
-
acquiring an image;
reducing the image with a Foreated Laplacian pyramid algorithm to form a plurality of down-sampled images, the down-sampled images decreasing in resolution from a highest to a lowest resolution;
designating a subset of each of the plurality of down-sampled images to form a plurality of subset images;
expanding the plurality of subset images to form a plurality of expanded subset images;
subtracting the plurality of expanded subset images from the plurality of down-sampled images to form a plurality of difference images; and
processing the lowest resolution down-sampled image with the plurality of difference images to form the compressed image. - View Dependent Claims (10, 11, 12, 13)
thresholding the plurality of difference images to form a plurality of threshold images;
quantizing the plurality of threshold images;
compressing the plurality of threshold images using a lossless compression algorithm to form a plurality of compression images; and
transmitting the lowest resolution down-sampled image with the plurality of compression images.
-
-
11. The method of claim 9, wherein the designating corresponds to one or more fixation points.
-
12. The method of claim 11, wherein the pointing device comprises a mouse.
-
13. The method of claim 9, wherein the designating corresponds one or more screen coordinates of a pointing device.
Specification