Image analyzing device using adaptive criteria

• US 5,825,936 A
• Filed: 10/15/1996
• Issued: 10/20/1998
• Est. Priority Date: 09/22/1994
• Status: Expired due to Term

• Current Assignee: University of South Florida (State University System of Florida)
• Original Assignee: University of South Florida (State University System of Florida)
• Inventors: Qian, Wei, Kallergi, Maria, Clarke, Laurence P.
• Primary Examiner(s): Mancuso, Joseph
• Assistant Examiner(s): Patel, Jayanti K.
• Application Number:
  US08/730,574
• Time in Patent Office: 735 Days
• Field of Search: 382/128, 382/129, 382/130, 382/131, 382/132, 382/133, 382/260, 382/261, 382/262, 382/263, 382/264, 382/265, 364/413.13
• US Class Current: 382/261
• CPC Class Codes: G06T 2207/20084 Artificial neural networks [ANN]
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 5/94 based on local image properties, e.g. for local contrast enhancement
• CPC Class Codes: G06T 7/0012 Biomedical image inspection

Abstract

A hybrid filter architecture and an artificial neural network is proposed for image enhancement and detection of suspicious areas in digital x-ray images that is operator, image, and digital x-ray sensor independent. The hybrid filter architecture includes an Adaptive Multistage Nonlinear Filter (AMNF) cascaded with an M-channel Tree Structured Wavelet Transform (TSWT). The AMNF shares the advantages of an array of linear and nonlinear filters and is adaptively supervised using either an order statistic or linear operator. The filter is used for noise suppression and image enhancement and adapts to the noise characteristics of the sensor. The TSWT is used for multiresolution image decomposition and reconstruction of subimages for further image enhancement of diagnostic features of interest. A Multistage Artificial Neural Network (MANN) is proposed, together with Kalman Filtering for network training, for both improved detection or classification of suspicious areas and computational efficiency to allow the MANN to be applied to full digital images without operator input. The hybrid filter architecture and MANN may be applied to any gray scale image in medical imaging. The specific application of the proposed method includes: (a) improved enhancement or detection of suspicious areas as a "second opinion" strategy using a computer workstation or (b) mass screening of large image databases such as that used for medical screening programs.

Claims

1. An imaging analyzing device for analyzing a digital image having at least one suspicious target area, normal parenchymal tissue and image background noise, the image analyzing device comprising:

first stage filter means for producing a first enhanced image, said first stage filter means including an adaptive multi-stage two dimensional using localized metrics as an adaptive criteria for filter type selection filter for filtering out background noise and selectively smoothing different tissues on a pixel by pixel basis in a manner to selectively enhance the suspicious target areas with the digital image; and

second stage filter means for producing a second stage enhanced image with selective enhancement of suspicious target areas including a decomposer means for multiresolution wavelet transform, for decomposing said first enhanced image into subimages and a reconstructor means for reconstructing selected subimages relating to possible suspicious target area and background tissues, summing the image representing the reconstructed sub-images representing the suspicious target areas with said first stage enhanced image, substracting the image representing the reconstructed sub-images representing the background tissues, including the use of appropriate weighting factors.

2. A device as set forth in claim 1, further characterized by said adaptive multi-stage two dimensional filter further including two dimensional linear, non-linear and alpha trimmed mean filters, and having different filter windows, providing different levels of smoothing of image pixels.

3. A device as set forth in claim 2, further characterized by said adaptive multi-stage filter further including an adaptive order statistic operation for texture enhancement of the suspicious target areas of the digital image.

4. A device as set forth in claim 2, further characterized by said adaptive multi-stage two dimensional filter further including an adaptive linear operation to enhance the contrast of the suspicious target areas of the digital image.

5. The device as set forth in any of claims 1, wherein the suspicious target areas further include possible microcalcifications and wherein said first and second stage filter means are designed to selectively enhance microcalcifications.

6. A device as set forth in claim 5, further characterized by including a neural network means for receiving said second stage enhanced image and extracting computed features to classify the suspicious target areas as either microcalcifications or normal tissues with the digital image.

7. A device as set forth in claim 6, further characterized by said neural network including a first and second stages, wherein said image pixels from the enhanced image means are used as input features to said first stage of said neural network, and wherein the output of said first stage of said neural network and additional extracted features from the original image are used as input to said second stage to classify the suspicious areas as either microcalcifications or normal tissue.