← Back to Patent Explorer
US5289548A Expired / Lapsed US ⏱️ 0.0 years remaining

Compression and reconstruction of radiological images

Assignee: Lockheed Martin Corp
•
Filing: 1992-06-30
•
Granted: 1994-02-22
•
Expires: 2012-06-30
📄 Original PDF 🌐 Google Patents 🔄 Sync Now

Abstract

Abstract The present invention provides for a compression scheme tailored to the compression of large radiological images processed with image processing workstations. The compression scheme is a variation of the CCITT JPEG compression scheme with special care taken to suppress blocking effects of the 8×8 discrete cosine transform used therein. In accordance with the present invention, an original image is downsampled, then interpolated back to its original size, resulting in a smoothed image. The difference between the smoothed image and the original image is then compressed using an 8×8 discrete cosine transform with 12 bit data. Major artifacts present in the original image are due to the mismatching of the low frequency components of the image at the edges of the blocks. When the low frequencies are removed by subtracting out the low frequency components, the artifacts disappear except in areas having very large dynamic changes. The compression scheme of the present invention is particularly useful for archiving or telecommunicating images, achieving compression of more than 10:1 without loosing the characteristics of the radiological image that permit effective diagnosis. A reconstruction scheme is also disclosed which allows recovery of the image from a compressed image file.

Inventors

  • 👤 Dennis L. Wilson
  • 👤 James T. Stenstrom

Bibliographic Details

Application Number —
Jurisdiction US
Data Source google_patents
Last Synchronized 2026-10-24 15:38
CPC Classifications
H; H04; H04N; H04N19/00; H04N19/85; H04N19/86; H04N19/10; H04N19/102; H04N19/124; H04N19/126; H04N19/134; H04N19/154; H04N19/30; H04N19/50; H04N19/503; H04N19/51; H04N19/527; H04N19/60

Claims Analysis

Total Claims: 22
Claim 1
Independent Claim 156 words
An image compression method comprising the steps of: downsampling an original radiological image to produce a downsampled image; interpolating the downsampled image back to its original size to provide a smoothed image; processing the smoothed image and the original image to provide a difference image; processing the difference image using a discrete cosine transform to provide a transformed difference image; frequency weighting the transformed difference image to provide a weighted difference image; quantizing the weighted difference image to provide a quantized difference image; encoding the quantized difference image to produce an encoded difference image; differential pulse code modulating the downsampled image to produce a differential pulse code modulated downsampled image; and encoding the differential pulse code modulated downsampled image to produce an encoded downsampled image; whereby the encoded difference image and the encoded downsampled image comprise a compressed image, and wherein compression is adapted to remove artifacts present in the original radiological image caused by blocking.
Claim 2
Dependent Claim 156 words
An image compression method comprising the steps of: downsampling an original radiological image to produce a downsampled image; interpolating the downsampled image back to its original size to provide a smoothed image; processing the smoothed image and the original image to provide a difference image; processing the difference image using a discrete cosine transform to provide a transformed difference image; frequency weighting the transformed difference image to provide a weighted difference image; quantizing the weighted difference image to provide a quantized difference image; encoding the quantized difference image to produce an encoded difference image; differential pulse code modulating the downsampled image to produce a differential pulse code modulated downsampled image; and encoding the differential pulse code modulated downsampled image to produce an encoded downsampled image; whereby the encoded difference image and the encoded downsampled image comprise a compressed image, and wherein compression is adapted to remove artifacts present in the original radiological image caused by blocking.
Claim 3
Dependent Claim 86 words
The method of claim 1 which further comprises the step of normalizing the compressed image by means of the steps of: finding the sum of the squares of the coefficients of the discrete cosine transform; determining a gain factor such that energy in the transformed difference image is equal to a constant; and applying the gain factor to each of the coefficients of the discrete cosine transform, changing the energy in the transformed difference image to a constant value that is equal to a normalization energy.
Claim 4
Dependent Claim 27 words
The method of claim 1 wherein the step of quantizing the weighted difference image comprises the step of rounding coefficients to quantize them to a nearest integer.
Claim 5
Dependent Claim 47 words
The method of claim 1 wherein, if the energy in the transformed difference image is larger that a predetermined threshold, the step of quantizing the weighted difference image comprises quantizing the coefficients at their original level so that the quantization noise is reduced, thus suppressing blocking effects.
Claim 6
Dependent Claim 25 words
The method of claim 1 wherein the step of encoding the quantized image comprises the step of encoding weighted coefficient values using a Huffman code.
Claim 7
Dependent Claim 46 words
The method of claim 1 wherein the step of encoding the quantized image comprises the step of encoding weighted coefficient values using an arithmetic code that codes the weighted, quantized coefficients in a code word that maps coefficient values into an interval from zero to one.
Claim 8
Dependent Claim 36 words
The method of claim 1 wherein the step of processing the difference image using a discrete cosine transform comprises the step of frequency weighting the transform coefficients wherein lower weights are given to higher frequency components.
Claim 9
Dependent Claim 37 words
The method of claim 1 wherein the step of processing the difference image comprises the steps of: deleting high frequency components having energy below a predetermined threshold; and Huffman coding the remaining quantized discrete cosine transform coefficients.
Claim 10
Dependent Claim 35 words
The method of claim 8 wherein the level of quantization of the discrete cosine transform coefficients is adjusted by setting a quantization level that is a function of the energy in the transformed difference image.
Claim 11
Dependent Claim 129 words
An image reconstruction method for use with a compressed image comprising an encoded difference radiological image and an encoded downsampled radiological image that produces a reconstructed image, said method comprising the steps of: decoding the encoded difference image to produce a decoded difference image; frequency weighting and removing the normalization of the decoded difference image to produce a restored transformed difference image; transforming the restored transformed difference image using an inverse discrete cosine transform to produce a restored difference image; decoding the encoded downsampled radiological image to produce a decoded image; differential pulse code demodulating the decoded image to produce a demodulated image; interpolating the demodulated image to provide a full scale smoothed image; and combining the compressed image and the smoothed image to produce a reconstructed original image.
Claim 12
Dependent Claim 129 words
An image reconstruction method for use with a compressed image comprising an encoded difference radiological image and an encoded downsampled radiological image that produces a reconstructed image, said method comprising the steps of: decoding the encoded difference image to produce a decoded difference image; frequency weighting and removing the normalization of the decoded difference image to produce a restored transformed difference image; transforming the restored transformed difference image using an inverse discrete cosine transform to produce a restored difference image; decoding the encoded downsampled radiological image to produce a decoded image; differential pulse code demodulating the decoded image to produce a demodulated image; interpolating the demodulated image to provide a full scale smoothed image; and combining the compressed image and the smoothed image to produce a reconstructed original image.
Claim 13
Dependent Claim 170 words
Image compression apparatus comprising: means for downsampling an original radiological image to produce a downsampled image; means for interpolating the downsampled image back to its original size to provide a smoothed image; means for processing the smoothed image and the original image to provide a difference image; means for processing the difference image using a discrete cosine transform to provide a transformed difference image; means for frequency weighting the transformed difference image to provide a weighted difference image; means for quantizing the weighted difference image to provide a quantized difference image; means for encoding the quantized difference image to produce an encoded difference image; means for differential pulse code modulating the downsampled image to produce a differential pulse code modulated downsampled image; and means for encoding the differential pulse code modulated downsampled image to produce an encoded downsampled image; whereby the encoded difference image and the encoded downsampled image comprise a compressed image, and wherein compression is adapted to remove artifacts present in the original radiological image caused by blocking.
Claim 14
Dependent Claim 170 words
Image compression apparatus comprising: means for downsampling an original radiological image to produce a downsampled image; means for interpolating the downsampled image back to its original size to provide a smoothed image; means for processing the smoothed image and the original image to provide a difference image; means for processing the difference image using a discrete cosine transform to provide a transformed difference image; means for frequency weighting the transformed difference image to provide a weighted difference image; means for quantizing the weighted difference image to provide a quantized difference image; means for encoding the quantized difference image to produce an encoded difference image; means for differential pulse code modulating the downsampled image to produce a differential pulse code modulated downsampled image; and means for encoding the differential pulse code modulated downsampled image to produce an encoded downsampled image; whereby the encoded difference image and the encoded downsampled image comprise a compressed image, and wherein compression is adapted to remove artifacts present in the original radiological image caused by blocking.
Claim 15
Dependent Claim 86 words
The apparatus of claim 11 which further comprises means for normalizing the compressed image comprising: means for finding the sum of the squares of the coefficients of the discrete cosine transform; means for determining a gain factor such that energy in the transformed difference image is equal to a constant; and means for applying the gain factor to each of the coefficients of the discrete cosine transform, changing the energy in the transformed difference image to a constant value that is equal to its normalization energy.
Claim 16
Dependent Claim 26 words
The apparatus of claim 11 wherein the means for quantizing the weighted difference image comprises means for rounding coefficients to quantize them to a nearest integer.
Claim 17
Dependent Claim 49 words
The apparatus of claim 11 wherein, if the energy in the transformed difference image is larger that a predetermined threshold, the means for quantizing the weighted difference image comprises means for quantizing the coefficients at their original level so that the quantization noise is reduced, thus suppressing blocking effects.
Claim 18
Dependent Claim 24 words
The apparatus of claim 11 wherein the means for encoding the quantized image comprises means for encoding weighted coefficient values using a Huffman code.
Claim 19
Dependent Claim 44 words
The apparatus of claim 11 wherein the means for coding the quantized image comprises means for encoding weighted coefficient values using an arithmetic code that codes weighted, quantized coefficients in a code word that maps coefficient values into an interval from zero to one.
Claim 20
Dependent Claim 35 words
The apparatus of claim 11 wherein the means for processing the difference image using a discrete cosine transform comprises means for frequency weighting the transform coefficients wherein lower weights are given to higher frequency components.
Claim 21
Dependent Claim 43 words
The apparatus of claim 11 wherein the means for processing the difference image comprises: means for deleting high frequency components containing energy below a predetermined threshold; and means for Huffman coding quantized discrete cosine transform coefficients remaining after deleting the high frequency components.
Claim 22
Dependent Claim 35 words
The apparatus of claim 18 wherein the level of quantization of the discrete cosine transform coefficients is adjusted by setting a quantization level that is a function of the energy in the transformed difference image.

References Cited (Backward Citations)

US4745474A Cited
US4774574A Cited
US5109451A Cited
US4939645A Cited
US5170264A Cited
US5046121A Cited
US5063608A Cited

USPTO Annuity & Maintenance Windows

Annuity Stage Window Opens Payment Deadline Surcharge Period Until Estimated Fee Status
3.5 Years (Year 4) 1997-02-21 1997-08-23 1998-02-22 $2000.00 Window Closed / Lapsed
7.5 Years (Year 8) 2001-02-21 2001-08-23 2002-02-22 $3760.00 Window Closed / Lapsed
11.5 Years (Year 12) 2005-02-21 2005-08-23 2006-02-22 $7700.00 Window Closed / Lapsed

Assigned Categories & Tags

No categories assigned to this patent yet.

Assign Category Tag