Prostate Segmentation From 3D Mri Using A Two-Stage Model and Variable-Input Based Uncertainty Measure

Abstract

This paper proposes a two-stage segmentation model, variable-input based uncertainty measures and an uncertainty-guided post-processing method for prostate segmentation on 3D magnetic resonance images (MRI). The two-stage model was based on 3D dilated U-Nets with the first stage to localize the prostate and the second stage to obtain an accurate segmentation from cropped images. For data augmentation, we proposed the variable-input method which crops the region of interest with additional random variations. Similar to other deep learning models, the proposed model also faced the challenge of suboptimal performance in certain testing cases due to varied training and testing image characteristics. Therefore, it is valuable to evaluate the confidence and performance of the network using uncertainty measures, which are often calculated from the probability maps or their standard deviations with multiple model outputs for the same testing case. However, few studies have quantitatively compared different methods of uncertainty calculation. Furthermore, unlike the commonly used Bayesian dropout during testing, we developed uncertainty measures based on the variable input images at the second stage and evaluated its performance by calculating the correlation with ground-truth-based performance metrics, such as Dice score. For performance estimation, we predicted Dice scores and Hausdorff distance with the most correlated uncertainty measure. For postprocessing, we performed Gaussian filter on the underperformed slices to improve segmentation quality. Using PROMISE-12 data, we demonstrated the robustness of the two-stage model and showed high correlation of the proposed variable-input based uncertainty measures with GT-based performance. The uncertainty-guided postprocessing method significantly improved label smoothness.

A review of medical image data augmentation techniques for deep learning applications

“…If an image has one or more contours associated with it, the same transformation is applied to the contours. Geometric transformations are so common that they were utilised by 92 of the 93 basic augmentation studies 15–106 …”

“…The two most commonly used splines in the literature reviewed were B‐splines (demonstrated in Fig. 4c) and thin plate splines, with 9 articles employing such methods 23,39,44,45,92,95,118–120 …”

“…4c) and thin plate splines, with 9 articles employing such methods. 23,39,44,45,92,95,[118][119][120] Krivov et al 26 developed a method wherein brain lesions could be mapped onto healthy patient scans, by performing deformable image registration between the original patient and a healthy one. Nalepa et al 121 demonstrated an approach using diffeomorphic image registration, co-registering pairs of lesion images to create augmented data which improved the generalisation of their DL models when combined with an affine augmentation.…”

Multi‐scale discriminative network for prostate cancer lesion segmentation in multiparametric MR images

“…Some researchers have proposed improved methods for prostate segmentation from multiple perspectives, such as designing additional structures or proposing new loss functions. Pan et al 22 . used two continuous U‐Net networks for coarse and fine segmentation to address the imbalance between the front and the background in prostate MR images.…”

An Optimized Approach for Prostate Image Segmentation Using K‐Means Clustering Algorithm with Elbow Method

“…Pan et al [24] proposed a prostate segmentation model on the 3D magnetic resonance images (MRIs). e model contains two stages: variable input-based uncertainty measures and an uncertainty-guided postprocessing method.…”