Computer-aided method for image feature analysis and diagnosis in mammography

• US 5,598,481 A
• Filed: 09/29/1995
• Issued: 01/28/1997
• Est. Priority Date: 04/29/1994
• Status: Expired due to Fees

• Current Assignee: ARCH Development Corporation (University of Chicago)
• Original Assignee: ARCH Development Corporation (University of Chicago)
• Inventors: Yoshida, Hiroyuki, Ema, Takehiro, Nishikawa, Robert M., Doi, Kunio
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): Tran, Phuoc
• Application Number:
  US08/536,450
• Time in Patent Office: 487 Days
• Field of Search: 382/130, 382/132, 382/254, 382/263, 382/264, 382/275, 382/276, 382/280, 382/281, 364/413.23
• US Class Current: 382/130
• CPC Class Codes: A61B 6/5258 involving detection or reduction of artifacts or noise
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/20064 Wavelet transform [DWT]
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06V 10/52 Scale-space analysis, e.g. wavelet analysis multi-scale boundary representations G06V10/42
• CPC Class Codes: Y10S 128/925 Neural network

Abstract

A method for automated detection of abnormal anatomic regions, wherein a mammogram is digitized to produce a digital image and the digital image is processed using local edge gradient analysis and linear pattern analysis in addition to feature extraction routines to identify abnormal anatomic regions. Noise reduction filtering and pit-filling/spike-removal filtering techniques are also provided. Multiple difference imaging techniques are also used in which difference images employing different filter characteristics are obtained and processing results logically OR'"'"'ed to identify abnormal anatomic regions. In another embodiment the processing results with and without noise reduction filtering are logically AND'"'"'ed to improve detection sensitivity. Also, in another embodiment the wavelet transform is utilized in the identification and detection of abnormal regions. The wavelet transform is preferably used in conjunction with the difference imaging technique with the results of the two techniques being logically OR'"'"'ed.

Claims

1. A method for automated detection of an abnormal anatomic region, comprising:

obtaining a digital image of an object including said anatomic region;

first filtering said digital image using a first set of filters having a first transfer function to produce a difference image based on a difference between a signal-enhanced and a signal suppressed image of said digital image;

second filtering said digital image using a wavelet decomposition and reconstruction of said digital image to produce a transformation image;

processing said difference image to identify in said difference image first locations which corresponds to potential abnormal regions;

processing said transformation image to identify in said transformation image second locations which correspond to potential abnormal regions;

logically OR'"'"'ing the first and second locations to identify as candidate abnormal regions all those first and second locations which are separated by a predetermined distance or more; and

,

processing the candidate abnormal regions to identify abnormal regions from among said candidate abnormal regions.

2. The method of claim 1, wherein:

said first filtering step comprises spatially filtering using a first set of spatial filters having a first optical transfer factor; and

said second filtering step comprises decomposing said digital image to the wavelet domain and reconstructing said digital image based on second and third level components in the decomposed digital image.

3. The method of claim 2, wherein said first filtering step comprises:

filtering using a spatial filter having an optical transfer factor in the frequency range of 1.0 to 1.5 cycles/mm.

4. The method of claim 3, wherein said steps of processing said difference image and processing said transformation image in each case comprise:

performing global thresholding;

performing local thresholding;

performing predetermined feature extraction routines to identify microcalcifications;

identifying microcalcification clusters based on the identified microcalcifications; and

identifying abnormal regions based on the identified microcalcification clusters.

5. The method of claim 1, further comprising:

determining an edge gradient for each of the candidate abnormal regions;

comparing each edge gradient determined in said determining step with at least one threshold; and

eliminating candidate abnormal regions identified in said processing step from consideration as an abnormal region based on a result of said comparing step.

6. The method of claim 5, wherein:

said comparing step comprises comparing each edge gradient with a predetermined number; and

said eliminating step comprises eliminating those candidate abnormal regions having an edge gradient exceeding said predetermined number.

7. The method according to claim 5, wherein:

said determining step comprises determining a degree of linearity for each of said locations identified;

said comparing step comprises comparing the degree to linearity determined for each location with a predetermined linearity threshold; and

said eliminating step comprises eliminating from consideration as abnormal regions locations identified in said processing step and having a linearity factor exceeding said predetermined linearity threshold.

8. The method of claim 5, wherein:

said comparing step comprises comparing each edge gradient with a varying threshold which varies inversely as a function of the average pixel value for said location; and

said eliminating step comprises eliminating those candidate abnormal regions having an edge gradient less than said varying threshold.

9. The method according to claim 6, wherein:

said determining step comprises determining a degree of linearity for each of said locations identified;

said comparing step comprises comparing the degree to linearity determined for each location with a predetermined linearity threshold; and

said eliminating step comprises eliminating from consideration as abnormal regions locations identified in said processing step and having a linearity factor exceeding said predetermined linearity threshold.

10. The method of claim 8, wherein said candidate abnormal region processing step comprises:

identifying locations of microcalcifications and locations of microcalcification clusters;

determining edge gradients for the locations of said microcalcifications and for the locations of said microcalcification clusters; and

comparing the edge gradients determined for said microcalcification locations and for said microcalcification cluster locations with respective thresholds; and

eliminating locations based on the comparing of microcalcification location edge gradients with respective thresholds and based on the comparing of microcalcification cluster edge gradients with respective thresholds.

11. The method according to claim 8, wherein:

said determining step comprises determining a degree of linearity for each of said locations identified;

said comparing step comprises comparing the degree to linearity determined for each location with a predetermined linearity threshold; and

said eliminating step comprises eliminating from consideration as abnormal regions locations identified in said processing step and having a linearity factor exceeding said predetermined linearity threshold.

12. The method according to claim 10, wherein:

said determining step comprises determining a degree of linearity for each of said locations identified;

said comparing step comprises comparing the degree to linearity determined for each location with a predetermined linearity threshold; and

said eliminating step comprises eliminating from consideration as abnormal regions locations identified in said processing step and having a linearity factor exceeding said predetermined linearity threshold.