Medical images are often used to detect and characterize pathology and disease; however, automatically identifying and segmenting pathology in medical images is challenging because the appearance of pathology across diseases varies widely. To address this challenge, we propose a Bayesian deep learning method that learns to translate healthy computed tomography images to magnetic resonance images and simultaneously calculates voxel-wise uncertainty. Since high uncertainty occurs in pathological regions of the image, this uncertainty can be used for unsupervised anomaly segmentation. We show encouraging experimental results on an unsupervised anomaly segmentation task by combining two types of uncertainty into a novel quantity we call scibilic uncertainty.
Image synthesis learns a transformation from the intensity features of an input image to yield a different tissue contrast of the output image. This process has been shown to have application in many medical image analysis tasks including imputation, registration, and segmentation. To carry out synthesis, the intensities of the input images are typically scaled—i.e., normalized—both in training to learn the transformation and in testing when applying the transformation, but it is not presently known what type of input scaling is optimal. In this paper, we consider seven different intensity normalization algorithms and three different synthesis methods to evaluate the impact of normalization. Our experiments demonstrate that intensity normalization as a preprocessing step improves the synthesis results across all investigated synthesis algorithms. Furthermore, we show evidence that suggests intensity normalization is vital for successful deep learning-based MR image synthesis.
Digital breast tomosynthesis (DBT) acquires a series of projection images from different angles as an x-ray source rotates around the breast. Such imaging geometry lends DBT naturally to stereoscopic viewing as two projection images with a reasonable separation angle can easily form a stereo pair. This simulation study assessed the efficacy of stereo viewing of DBT projection images. Three-dimensional computational breast phantoms with realistically shaped synthetic lesions were scanned by three simulated DBT systems. The projection images were combined into a sequence of stereo pairs and presented to a stereomatching-based model observer for deciding lesion presence. Signal-to-noise ratio was estimated, and the detection performance with stack viewing of reconstructed slices was the benchmark. We have shown that: (1) stereo viewing of projection images may underperform stack viewing of reconstructed slices for current DBT geometries; (2) DBT geometries may impact the efficacy of the two viewing modes differently: narrow-arc and wide-arc geometries may be better for stereo viewing and stack viewing, respectively; (3) the efficacy of stereo viewing may be more robust than stack viewing to reductions in dose. While in principle stereo viewing is potentially effective for visualizing DBT data, effective stereo viewing may require specifically optimized DBT image acquisition.
Proc. SPIE. 10136, Medical Imaging 2017: Image Perception, Observer Performance, and Technology Assessment
KEYWORDS: Signal to noise ratio, Breast, Data modeling, Sensors, Computing systems, Computer simulations, 3D modeling, Digital imaging, Signal detection, Performance modeling, 3D vision, 3D image processing, Digital breast tomosynthesis, Breast imaging
Stereoscopic views of 3D breast imaging data may better reveal the 3D structures of breasts, and potentially improve the detection of breast lesions. The imaging geometry of digital breast tomosynthesis (DBT) lends itself naturally to stereo viewing because a stereo pair can be easily formed by two projection images with a reasonable separation angle for perceiving depth. This simulation study attempts to mimic breast lesion detection on stereo viewing of a sequence of stereo pairs of DBT projection images. 3D anthropomorphic computational breast phantoms were scanned by a simulated DBT system, and spherical signals were inserted into different breast regions to imitate the presence of breast lesions. The regions of interest (ROI) had different local anatomical structures and consequently different background statistics. The projection images were combined into a sequence of stereo pairs, and then presented to a stereo matching model observer for determining lesion presence. The signal-to-noise ratio (SNR) was used as the figure of merit in evaluation, and the SNR from the stack of reconstructed slices was considered as the benchmark. We have shown that: 1) incorporating local anatomical backgrounds may improve lesion detectability relative to ignoring location-dependent image characteristics. The SNR was lower for the ROIs with the higher local power-law-noise coefficient β. 2) Lesion detectability may be inferior on stereo viewing of projection images relative to conventional viewing of reconstructed slices, but further studies are needed to confirm this observation.