Image processing method, system and apparatus for noise reduction in an image sequence representing a threadlike moving object

• US 6,574,300 B1
• Filed: 11/30/2000
• Issued: 06/03/2003
• Est. Priority Date: 03/30/1999
• Status: Expired due to Fees

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Florent, Raoul, Mequio, Claude
• Primary Examiner(s): Dunn, Drew A.
• Application Number:
  US09/701,630
• Time in Patent Office: 915 Days
• Field of Search: 378/4, 378/19, 378/901, 378/62, 378/98.12, 378/98, 382/132
• US Class Current: 378/19
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10121 Fluoroscopy
• CPC Class Codes: G06T 2207/20032 Median filtering
• CPC Class Codes: G06T 2207/20182 Noise reduction or smoothing in the temporal domain; Spatio-temporal filtering
• CPC Class Codes: G06T 2207/30021 Catheter; Guide wire
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: Y10S 378/901 Computer tomography program or processor

Abstract

The invention relates to an image processing method for noise reduction in a sequence of noisy original images representing a threadlike structure moving on a background, said method comprising: an acquisition of the digital pixel value at a current pixel location in a first and a second successive original image of the sequence, an extraction of the thread-like structure for providing first and second binary mask image pixel values at said location from said first and second original image, spatial and time-filtering of the pixel value of the second original image, and a pixel value selection based on said binary mask image pixel values for providing respectively, as a finally filtered pixel value, either the spatially or time-filtered pixel value respectively, as a threadlike structure filtered or a background pixel value. The invention also relates to a system and an apparatus for carrying out the method.

Claims

1. An image processing method for noise reduction in a sequence of noisy original images representing a thread-like structure moving on a background, said method comprising an acquisition of a pixel value at a current pixel location in a first and a second successive original image of the sequence, and comprising processing steps for producing a filtered pixel value at said current pixel location for reconstructing a filtered second image, wherein are included:

a step of extraction of the threadlike structure for producing first and second binary mask image pixel values at said location from said first and second original image pixel values, steps of producing spatially and temporally filtered pixel values of the second original image, and a step of pixel value selection based on said binary mask image pixel values for producing respectively, as a finally filtered pixel value, either the temporally filtered pixel value as a background filtered pixel value, or the spatially filtered pixel value as a threadlike-structure filtered pixel value.

2. A method as claimed in claim 1, wherein:

the step of pixel value selection determines, from the first and second binary mask image pixel values, whether the current pixel has been a background pixel in the two successive binary mask images, or has been a threadlike structure pixel value in at least one of the two successive binary mask images, for providing respectively, as a finally filtered pixel value, either the temporally filtered pixel value as a background filtered pixel value, or the spatially filtered pixel value as a threadlike-structure filtered pixel value.

3. A method as claimed in claim 2, wherein:

as the mask images provide binary pixel values, the step of pixel value selection comprises a switch operation for issuing either the temporally or the spatially filtered pixel value according to a binary control signal value based on said binary pixel values at the current pixel location in two successive mask images.

4. A method as claimed in claim 3, wherein:

the binary control signal value is produced by a logic OR operation whose inputs are the two successive binary mask image pixel values at the current pixel location and whose output is either a first binary control signal value (0) to trigger the temporally filtered pixel value when the two successive binary mask image pixel values both correspond to background or a second binary control signal value (1) to trigger the spatially filtered pixel value when at least one of the two successive binary mask image pixel values corresponds to the threadlike structure.

5. A method as claimed in claim 4, wherein:

the spatially filtered pixel value produced by the switch operation is enhanced by an enhancing operation based on the pixel value selection operation.

6. A method as claimed in claim 5, wherein:

in the mask images, the pixel values are binary signals which are given the zero (0) value for ground pixels and the one (1) value for thread-like structure pixels, and the first binary control signal which triggers the time-filtered pixel values corresponding to background has the zero value, whereas the second binary control signal which triggers the spatially filtered pixel values corresponding to the threadlike structure has the one value.

7. A method as claimed in one of claim 5, wherein:

the operation to enhance the spatially filtered value comprises a multiplying operation by a constant value (C) applied to the second binary mask image pixel value so as to produce either a zero (0) signal value or said constant value (C) respectively, depending on whether said second mask pixel value is a ground pixel or a thread-like structure pixel, and comprises an adding operation applied to the filtered pixel value issued by the switch operation, which subtracts the zero value (0) when the switch operation triggers the time-filtered pixel value corresponding to background, and which subtracts the constant value (C) when the switch operation triggers the spatially filtered pixel value corresponding to the threadlike-structure.

8. A method as claimed in claim 1, wherein:

the temporal filter is a recursive adaptive temporal filter and the spatial filter includes median filtering.

9. A system comprising a suitably programmed computer of a workstation or a special-purpose processor having circuit means, which are arranged to process image data according to the method for noise reduction in a sequence of noisy original images representing a thread-like structure moving on a background, said method comprising an acquisition of a pixel value at a current pixel location in a first and a second successive original image of the sequence, and comprising processing steps for producing a filtered pixel value at said current pixel location for reconstructing a filtered second image, wherein are included:

a step of extraction of the threadlike structure for producing first and second binary mask image pixel values at said location from said first and second original image pixel values, steps of producing spatially and temporally filtered pixel values of the second original image, and a step of pixel value selection based on said binary mask image pixel values for producing respectively, as a finally filtered pixel value, either the temporally filtered pixel value as a background filtered pixel value, or the spatially filtered pixel value as a threadlike-structure filtered pixel value, and having means for displaying images processed according to said method, and occasionally means for storing the image data.

10. An X-ray apparatus having means for acquiring digital medical image data and having a system comprising a suitably programmed computer of a workstation or a special-purpose processor having circuit means, which are arranged to process image data according to the method for noise reduction in a sequence of noisy original images representing a thread-like structure moving on a background, said method comprising an acquisition of a pixel value at a current pixel location in a first and a second successive original image of the sequence, and comprising processing steps for producing a filtered pixel value at said current pixel location for reconstructing a filtered second image, wherein are included:

a step of extraction of the threadlike structure for producing first and second binary mask image pixel values at said location from said first and second original image pixel values, steps of producing spatially and temporally filtered pixel values of the second original image, and a step of pixel value selection based on said binary mask image pixel values for producing respectively, as a finally filtered pixel value, either the temporally filtered pixel value as a background filtered pixel value, or the spatially filtered pixel value as a threadlike-structure filtered pixel value, and having means for displaying images processed according to said method, and occasionally means for storing the image data that has access to said digital medical image data for processing said image data.