AM-UNet: automated mini 3D end-to-end U-net based network for brain claustrum segmentation

Abstract

Recent advances in deep learning (DL) have provided promising solutions to medical image segmentation. Among existing segmentation approaches, the U-Net-based methods have been used widely. However, very few U-Net-based studies have been conducted on automatic segmentation of the human brain claustrum (CL). The CL segmentation is challenging due to its thin, sheet-like structure, heterogeneity of its image modalities and formats, imperfect labels, and data imbalance. We propose an automatic optimized U-Net-based 3D segmentation model, called AM-UNet, designed as an end-to-end process of the pre and post-process techniques and a U-Net model for CL segmentation. It is a lightweight and scalable solution which has achieved the state-of-the-art accuracy for automatic CL segmentation on 3D magnetic resonance images (MRI). On the T1/T2 combined MRI CL dataset, AM-UNet has obtained excellent results, including Dice, Intersection over Union (IoU), and Intraclass Correlation Coefficient (ICC) scores of 82%, 70%, and 90%, respectively. We have conducted the comparative evaluation of AM-UNet with other pre-existing models for segmentation on the MRI CL dataset. As a result, medical experts confirmed the superiority of the proposed AM-UNet model for automatic CL segmentation. The source code and model of the AM-UNet project is publicly available on GitHub: https://github.com/AhmedAlbishri/AM-UNET.

Fig. 1

The human claustrum (orange) delineated in a T1-weighted MRI

Fig. 2

AM-UNet Architecture

Fig. 3

Three Views of MRI Scans (from left to right) (a) Axial, (b) Coronal, (c) Sagittal

Fig. 4

T1w, T2w, and T1T2w MRIs and Normalized Pixel Intensities Graph

Fig. 5

Preprocessing for CL Segmentation: (1) 3D MRI Slicing (2) Slice Selection (3) ROI

Fig. 6

Postprocessing for 3D MRI Reconstruction: (1) Filtering False Positives, (2) Reconstruction of 2D Slices, (3) Reconstruction of 3D MRI

Fig. 7

Best and worst cases of AM-UNet

Fig. 8

Subject-wise Evaluation for 3D CL Segmentation: Dice Coefficient Scores (on left), IoU Scores (on right)

Fig. 9

AM-UNet for CL Segmentation for 3 subjects: (a) Ground Truth, (b) Ground Truth (CL ROI), (c) Predicted Outcome

Fig. 10

AM-UNet 3D Segmentation (Green) and Ground Truth (Orange)

Fig. 11

Evaluation with Dice Score & IoU Score: AM-UNet, U-Net, and UNet++

Fig. 12

Comparative Evaluation: Original and ROI Images, Ground Truth (white), AM-UNet (Green), U-Net (Red), and UNet++ (Blue)

Fig. 13

Comparative Analysis with Worst Cases of U-Net and UNet++