Unregistered Multiview Mammogram Analysis with Pre-trained Deep Learning Models
We show two important findings on the use of deep convolutional neural networks (CNN) in medical image analysis. First, we show that CNN models that are pre-trained using computer vision databases (e.g., Imagenet) are useful in medical image applications,
- PDF / 2,476,001 Bytes
- 9 Pages / 439.363 x 666.131 pts Page_size
- 36 Downloads / 246 Views
ACVT, University of Adelaide, Australia ISR, Instituto Superior Tecnico, Portugal 3 University of Queensland, Australia
Abstract. We show two important findings on the use of deep convolutional neural networks (CNN) in medical image analysis. First, we show that CNN models that are pre-trained using computer vision databases (e.g., Imagenet) are useful in medical image applications, despite the significant differences in image appearance. Second, we show that multiview classification is possible without the pre-registration of the input images. Rather, we use the high-level features produced by the CNNs trained in each view separately. Focusing on the classification of mammograms using craniocaudal (CC) and mediolateral oblique (MLO) views and their respective mass and micro-calcification segmentations of the same breast, we initially train a separate CNN model for each view and each segmentation map using an Imagenet pre-trained model. Then, using the features learned from each segmentation map and unregistered views, we train a final CNN classifier that estimates the patient’s risk of developing breast cancer using the Breast Imaging-Reporting and Data System (BI-RADS) score. We test our methodology in two publicly available datasets (InBreast and DDSM), containing hundreds of cases, and show that it produces a volume under ROC surface of over 0.9 and an area under ROC curve (for a 2-class problem - benign and malignant) of over 0.9. In general, our approach shows state-of-the-art classification results and demonstrates a new comprehensive way of addressing this challenging classification problem. Keywords: Deep learning, Mammogram, Multiview classification.
1
Introduction
Deep learning models are producing quite competitive results in computer vision and machine learning [1], but the application of such large capacity models in medical image analysis is complicated by the fact that they need large training sets that are rarely available in our field. This issue is circumvented in computer vision and machine learning with the use of publicly available pre-trained deep learning models, which are estimated with large annotated databases [2]
This work was partially supported by the Australian Research Council’s Discovery Projects funding scheme (project DP140102794). Prof. Bradley is the recipient of an Australian Research Council Future Fellowship(FT110100623).
c Springer International Publishing Switzerland 2015 N. Navab et al. (Eds.): MICCAI 2015, Part III, LNCS 9351, pp. 652–660, 2015. DOI: 10.1007/978-3-319-24574-4_78
Unregist. Multiview Mammogram Analysis with Pre-trained Deep Learning
653
Fig. 1. Model proposed in this paper using unregistered CC/MLO views and MC/Mass segmentations of the same breast, where the classification of the patient’s risk of developing breast cancer uses a CNN pre-trained with Imagenet [2].
and re-trained (or fine-tuned) for other problems that contain smaller annotated training sets. This fine-tuning process has been shown to improve the generalization of the model, compared to a mo
Data Loading...
