An approach to automated detection of tumors in mammograms

doi:10.1109/42.57760

Home
/
Papers
/
An approach to automated detection of tumors in mammograms

Journal Article•DOI•

An approach to automated detection of tumors in mammograms

D. Brzakovic¹, X.M. Luo², P. Brzakovic•Institutions (2)

University of Tennessee¹, Science Applications International Corporation²

01 Jan 1990-IEEE Transactions on Medical Imaging (IEEE)-Vol. 9, Iss: 3, pp 233-241

TL;DR: An automated system for detecting and classifying particular types of tumors in digitized mammograms is described, which uses a classification hierarchy to identify benign and malignant tumors.

read less

Abstract: An automated system for detecting and classifying particular types of tumors in digitized mammograms is described The analysis of mammograms is performed in two stages First, the system identifies pixel groupings that may correspond to tumors Next, detected pixel groupings are subjected to classification The essence of the first processing stage is multiresolution image processing based on fuzzy pyramid linking The second stage uses a classification hierarchy to identify benign and malignant tumors Each level of the hierarchy uses deterministic or Bayes classifiers and a particular measurement The measurements pertain to shape and intensity characteristics of particular types of tumors The classification hierarchy is organized in such a way that the simplest measurements are used at the top, with the system stepping through the hierarchy only when it cannot classify the detected pixel groupings with certainty >

...read moreread less

Citations

PDF

Open Access

More filters

Journal Article•DOI•

Approaches for automated detection and classification of masses in mammograms

[...]

Heng-Da Cheng¹, X. J. Shi¹, R. Min¹, Liming Hu¹, Xiaopeng Cai¹, H. N. Du¹ - Show less +2 more•Institutions (1)

Utah State University¹

01 Apr 2006-Pattern Recognition

TL;DR: The methods for mass detection and classification for breast cancer diagnosis are discussed, and their advantages and drawbacks are compared.

...read moreread less

526 citations

Cites background or methods from "An approach to automated detection ..."

...Global thresholding Global thresholding has been widely used for segmentation [39,42,43,46]....
[...]
...The classification hierarchy in [42] used the deterministic or Bayesian classifiers with four features to perform classification....
[...]
...A hierarchic classification example [42]....
[...]
...Binary decision tree [42,44,45,53,54,90] A binary decision tree recursively using a threshold to separate mammogram data into two classes each time Intensity features, shape features, texture features Low complexity Accuracy depends fully on the design of the decision tree and the features...
[...]
...[153,154] proposed a new cascade classifier ART2LDA combining an unsupervised classifier ART2 and a supervised classifier based on LDA to improve classification performance Hierarchical architecture Individual classifiers are combined into a tree structure, and each node is associated with a classifier In [42], the first level associated with deterministic classification and the area feature, and the rest three levels used Bayesian classifier and other types of features, such as shape descriptor, edge distance variation descriptor and edge intensity variation...
[...]

Journal Article•DOI•

Classification of mass and normal breast tissue: a convolution neural network classifier with spatial domain and texture images

[...]

Berkman Sahiner¹, Heang Ping Chan¹, Nicholas Petrick¹, Datong Wei², Mark A. Helvie¹, Dorit D. Adler¹, Mitchell M. Goodsitt¹ - Show less +3 more•Institutions (2)

University of Michigan¹, University of Chicago²

01 Oct 1996-IEEE Transactions on Medical Imaging

TL;DR: The authors' results demonstrate the feasibility of using a convolution neural network for classification of masses and normal tissue on mammograms using a generalized, fast and stable implementation of the CNN.

...read moreread less

Abstract: The authors investigated the classification of regions of interest (ROI's) on mammograms as either mass or normal tissue using a convolution neural network (CNN). A CNN is a backpropagation neural network with two-dimensional (2-D) weight kernels that operate on images. A generalized, fast and stable implementation of the CNN was developed. The input images to the CNN were obtained from the ROI's using two techniques. The first technique employed averaging and subsampling. The second technique employed texture feature extraction methods applied to small subregions inside the ROI. Features computed over different subregions were arranged as texture images, which were subsequently used as CNN inputs. The effects of CNN architecture and texture feature parameters on classification accuracy were studied. Receiver operating characteristic (ROC) methodology was used to evaluate the classification accuracy. A data set consisting of 168 ROIs containing biopsy-proven masses and 504 ROI's containing normal breast tissue was extracted from 168 mammograms by radiologists experienced in mammography. This data set was used for training and testing the CNN. With the best combination of CNN architecture and texture feature parameters, the area under the test ROC curve reached 0.87, which corresponded to a true-positive fraction of 90% at a false positive fraction of 31%. The authors' results demonstrate the feasibility of using a CNN for classification of masses and normal tissue on mammograms.

...read moreread less

414 citations

Journal Article•DOI•

A review of automatic mass detection and segmentation in mammographic images.

[...]

Arnau Oliver¹, Jordi Freixenet¹, Joan Martí¹, Elsa Pérez, Josep Pont, Erika R. E. Denton², Reyer Zwiggelaar³ - Show less +3 more•Institutions (3)

University of Girona¹, Norfolk and Norwich University Hospital², Aberystwyth University³

01 Apr 2010-Medical Image Analysis

TL;DR: The aim of this paper is to review existing approaches to the automatic detection and segmentation of masses in mammographic images, highlighting the key-points and main differences between the used strategies.

...read moreread less

388 citations

Cites background or methods from "An approach to automated detection ..."

...They demonstrated that this algorithm was directly correlated with the isodata clustering algorithm (Brzakovic et al., 1990)....
[...]
...Breast cancer remains, in the United States as well in the European Union, the leading cause of death for women after their 40 s (Eurostat, 2002; Buseman et al., 2003). However, although breast cancer incidence has increased over the past decade, breast cancer mortality has declined among women of all ages (Sickles, 1997). This favourable trend in mortality reduction may relate to the widespread adoption of mammography screening (Sickles, 1997; De Koning et al., 1995), and improvements made in breast cancer treatment (Buseman et al., 2003). Mammography remains the key screening tool for the detection of breast abnormalities. Vacek et al. (2002) show that the proportion of breast tumours that were detected in Vermont by mammographic screening increased from 2% during 1974–1984 to 36% during 1995–1999....
[...]
...One of the earliest approaches to mass detection was the work of Brzakovic et al. (1990), which was based on a multi-resolution fuzzy pyramid linking approach, a data structure in which the input image formed the basis of the pyramid and each subsequent level (of lower resolution) was sequentially…...
[...]

Journal Article•DOI•

A curated mammography data set for use in computer-aided detection and diagnosis research.

[...]

Rebecca Sawyer Lee¹, Francisco Gimenez¹, Assaf Hoogi¹, Kanae Miyake¹, Mia Gorovoy¹, Daniel L. Rubin¹ - Show less +2 more•Institutions (1)

Stanford University¹

19 Dec 2017-Scientific Data

TL;DR: The data set, the CBIS-DDSM (Curated Breast Imaging Subset of DDSM), includes decompressed images, data selection and curation by trained mammographers, updated mass segmentation and bounding boxes, and pathologic diagnosis for training data, formatted similarly to modern computer vision data sets.

...read moreread less

Abstract: Published research results are difficult to replicate due to the lack of a standard evaluation data set in the area of decision support systems in mammography; most computer-aided diagnosis (CADx) and detection (CADe) algorithms for breast cancer in mammography are evaluated on private data sets or on unspecified subsets of public databases. This causes an inability to directly compare the performance of methods or to replicate prior results. We seek to resolve this substantial challenge by releasing an updated and standardized version of the Digital Database for Screening Mammography (DDSM) for evaluation of future CADx and CADe systems (sometimes referred to generally as CAD) research in mammography. Our data set, the CBIS-DDSM (Curated Breast Imaging Subset of DDSM), includes decompressed images, data selection and curation by trained mammographers, updated mass segmentation and bounding boxes, and pathologic diagnosis for training data, formatted similarly to modern computer vision data sets. The data set contains 753 calcification cases and 891 mass cases, providing a data-set size capable of analyzing decision support systems in mammography.

...read moreread less

387 citations

Journal Article•DOI•

Mammographic feature enhancement by multiscale analysis

[...]

Andrew F. Laine¹, Sergio Schuler¹, Jian Fan¹, Walter Huda•Institutions (1)

University of Florida¹

01 Dec 1994-IEEE Transactions on Medical Imaging

TL;DR: It is demonstrated that features extracted from multiresolution representations can provide an adaptive mechanism for accomplishing local contrast enhancement and by improving the visualization of breast pathology, one can improve chances of early detection while requiring less time to evaluate mammograms for most patients.

...read moreread less

Abstract: Introduces a novel approach for accomplishing mammographic feature analysis by overcomplete multiresolution representations. The authors show that efficient representations may be identified within a continuum of scale-space and used to enhance features of importance to mammography. Methods of contrast enhancement are described based on three overcomplete multiscale representations: 1) the dyadic wavelet transform (separable), 2) the /spl phi/-transform (nonseparable, nonorthogonal), and 3) the hexagonal wavelet transform (nonseparable). Multiscale edges identified within distinct levels of transform space provide local support for image enhancement. Mammograms are reconstructed from wavelet coefficients modified at one or more levels by local and global nonlinear operators. In each case, edges and gain parameters are identified adaptively by a measure of energy within each level of scale-space. The authors show quantitatively that transform coefficients, modified by adaptive nonlinear operators, can make more obvious unseen or barely seen features of mammography without requiring additional radiation. The authors' results are compared with traditional image enhancement techniques by measuring the local contrast of known mammographic features. They demonstrate that features extracted from multiresolution representations can provide an adaptive mechanism for accomplishing local contrast enhancement. By improving the visualization of breast pathology, one can improve chances of early detection while requiring less time to evaluate mammograms for most patients. >

...read moreread less

382 citations

Cites methods from "An approach to automated detection ..."

...al [14] developed an automated system for the detection and classi cation of particular types of tumors in digitized mammograms....
[...]

1
2
3
4
…
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61

Collapse

References

PDF

Open Access

More filters

Journal Article•DOI•

A threshold selection method from gray level histograms

[...]

Nobuyuki Otsu

01 Jan 1979-IEEE Transactions on Systems, Man, and Cybernetics

37,017 citations

"An approach to automated detection ..." refers methods in this paper

...Next, the output is thresholded using algorithm described in [ 22 ], and possible tumors (that always have higher intensity) are assigned intensity 255, while the background is assigned intensity 0. The next processing stage is to determine if the detected regions are tumors, and, furthermore, to identify the benign tumors....
[...]

Journal Article•DOI•

Theory of Edge Detection

[...]

David Marr¹, Ellen C. Hildreth¹•Institutions (1)

Massachusetts Institute of Technology¹

29 Feb 1980-Proceedings of The Royal Society B: Biological Sciences

TL;DR: The theory of edge detection explains several basic psychophysical findings, and the operation of forming oriented zero-crossing segments from the output of centre-surround ∇2G filters acting on the image forms the basis for a physiological model of simple cells.

...read moreread less

Abstract: A theory of edge detection is presented. The analysis proceeds in two parts. (1) Intensity changes, which occur in a natural image over a wide range of scales, are detected separately at different scales. An appropriate filter for this purpose at a given scale is found to be the second derivative of a Gaussian, and it is shown that, provided some simple conditions are satisfied, these primary filters need not be orientation-dependent. Thus, intensity changes at a given scale are best detected by finding the zero values of delta 2G(x,y)*I(x,y) for image I, where G(x,y) is a two-dimensional Gaussian distribution and delta 2 is the Laplacian. The intensity changes thus discovered in each of the channels are then represented by oriented primitives called zero-crossing segments, and evidence is given that this representation is complete. (2) Intensity changes in images arise from surface discontinuities or from reflectance or illumination boundaries, and these all have the property that they are spatially. Because of this, the zero-crossing segments from the different channels are not independent, and rules are deduced for combining them into a description of the image. This description is called the raw primal sketch. The theory explains several basic psychophysical findings, and the operation of forming oriented zero-crossing segments from the output of centre-surround delta 2G filters acting on the image forms the basis for a physiological model of simple cells (see Marr & Ullman 1979).

...read moreread less

6,893 citations

"An approach to automated detection ..." refers methods in this paper

...Consequently, the regions around the boundary of the extracted region are subjected to local edge detection, using the Marr-Hildreth operator [ 20 ],...
[...]

Book•

Pattern recognition principles

[...]

Julius T. Tou¹, Rafael C. Gonzalez²•Institutions (2)

University of Florida¹, University of Tennessee²

01 Jan 1974

TL;DR: The present work gives an account of basic principles and available techniques for the analysis and design of pattern processing and recognition systems.

...read moreread less

Abstract: The present work gives an account of basic principles and available techniques for the analysis and design of pattern processing and recognition systems. Areas covered include decision functions, pattern classification by distance functions, pattern classification by likelihood functions, the perceptron and the potential function approaches to trainable pattern classifiers, statistical approach to trainable classifiers, pattern preprocessing and feature selection, and syntactic pattern recognition.

...read moreread less

3,237 citations

Journal Article•DOI•

A Convergence Theorem for the Fuzzy ISODATA Clustering Algorithms

[...]

James C. Bezdek¹•Institutions (1)

Utah State University¹

01 Jan 1980-IEEE Transactions on Pattern Analysis and Machine Intelligence

TL;DR: The theory of Zangwill is used to prove that arbitrary sequences generated by these (Picard iteration) procedures always terminates at a local minimum, or at worst, always contains a subsequence which converges to aLocal minimum of the generalized least squares objective functional which defines the problem.

...read moreread less

Abstract: In this paper the convergence of a class of clustering procedures, popularly known as the fuzzy ISODATA algorithms, is established. The theory of Zangwill is used to prove that arbitrary sequences generated by these (Picard iteration) procedures always terminates at a local minimum, or at worst, always contains a subsequence which converges to a local minimum of the generalized least squares objective functional which defines the problem.

...read moreread less

965 citations

Book Chapter•DOI•

The Pyramid as a Structure for Efficient Computation

[...]

P. J. Burt¹•Institutions (1)

Rensselaer Polytechnic Institute¹

01 Jan 1984

TL;DR: Here the pyramid will be viewed primarily as a computational tool, however, interesting similarities will be noted between pyramid processing and processing within the human visual system.

...read moreread less

Abstract: Many basic image operations may be performed efficiently within pyramid structures. Pyramid algorithms can generate sets of low-and band-pass filtered images at a fraction of the cost of the FFT. Local image properties such as texture statistics can be estimated with equal efficiency within Gaussianlike windows of many sizes. Pyramids support fast “coarse-fine” search strategies. Pyramids also provide a neural-like image representation which is robust, compact and appropriate for a variety of higher level tasks including motion analysis. Through “linking,” pyramids may be used to isolate and represent image segments of arbitrary size and shape. Here the pyramid will be viewed primarily as a computational tool. However, interesting similarities will be noted between pyramid processing and processing within the human visual system.

...read moreread less

452 citations