Breast density pattern characterization by histogram features and texture descriptors

Abstract

Introduction: Breast cancer is the first leading cause of death for women in Brazil as well as in most countries in the world. Due to the relation between the breast density and the risk of breast cancer, in medical practice, the breast density classification is merely visual and dependent on professional experience, making this task very subjective. The purpose of this paper is to investigate image features based on histograms and Haralick texture descriptors so as to separate mammographic images into categories of breast density using an Artificial Neural Network. Methods: We used 307 mammographic images from the INbreast digital database, extracting histogram features and texture descriptors of all mammograms and selecting them with the K-means technique. Then, these groups of selected features were used as inputs of an Artificial Neural Network to classify the images automatically into the four categories reported by radiologists. Results: An average accuracy of 92.9% was obtained in a few tests using only some of the Haralick texture descriptors. Also, the accuracy rate increased to 98.95% when texture descriptors were mixed with some features based on a histogram. Conclusion: Texture descriptors have proven to be better than gray levels features at differentiating the breast densities in mammographic images. From this paper, it was possible to automate the feature selection and the classification with acceptable error rates since the extraction of the features is suitable to the characteristics of the images involving the problem.

Breast Density Classification with Deep Convolutional Neural Networks

“…Table 2 in [16] for a comprehensive review) for this task involve two separate steps of feature extraction and classification. Carneiro et al [17] [18] and Fonseca et al [19] apply convolutional neural network too, however, only to extract features for another classification model (MLP and SVM respectively). Additionally, their results are obtained with a much smaller dataset (less than one thousand exams) of low resolution images.…”

Detection of partially overlapped masses in mammograms

“…In general, the radiologist can recognize most of tumors in the mammograms especially in the case of fatty tissues. Unfortunately, other types with high density are elusive and hard to be detected [11][12][13]. The difference between fatty and dense tissues mammograms can be easily viewed as shown in Figure 1.…”

PHOC Descriptor Applied for Mammography Classification

“…Carneiro et al [28] describe an approach conceptually similar to Bag of Features (a step in PHOW determination [19]) classifying mammographic images based on breast density using an Artificial Neural Network (ANN). They extracted features and applied K-means for determine the best feature for each class of mammography, then trained a ANN for classification.…”

“…The breast is composed by different tissues: milk glands, milk ducts and the breast density can vary due to hormonal factors [28]. This plethora of elements influences in image formation based on RX producing a complex image.…”

“…(b) Neuron activation function: log-sigmoid [28] (c) Training function: although Carneiro et al [28] used gradient descent with momentum and adaptive learning rate backpropagation, here was applied scaled conjugate gradient backpropagation as training function;…”