FU-Net: Multi-class Image Segmentation Using Feedback Weighted U-Net

Jafari, Mina; Li, Ruizhe; Xing, Yue; Auer, Dorothee P.; Francis, Susan T.; Garibaldi, Jonathan M.; Chen, Xin

doi:10.1007/978-3-030-34110-7_44

Cited by 11 publications

(7 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Multiclass segmentation has been considered challenging since it faces class data imbalance and interclass feature similarity problems (Chen et al 2018; Novikov et al 2018; Jafari et al 2019). As compared with multiclass strategies, binary strategies are generally more robust and achieve higher accuracy but come at the cost of increased training time (Berstad et al 2018; Gómez et al 2020).…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Multiclass CBCT Image Segmentation for Orthodontics with Deep Learning

et al. 2021

View full text Add to dashboard Cite

Accurate segmentation of the jaw (i.e., mandible and maxilla) and the teeth in cone beam computed tomography (CBCT) scans is essential for orthodontic diagnosis and treatment planning. Although various (semi)automated methods have been proposed to segment the jaw or the teeth, there is still a lack of fully automated segmentation methods that can simultaneously segment both anatomic structures in CBCT scans (i.e., multiclass segmentation). In this study, we aimed to train and validate a mixed-scale dense (MS-D) convolutional neural network for multiclass segmentation of the jaw, the teeth, and the background in CBCT scans. Thirty CBCT scans were obtained from patients who had undergone orthodontic treatment. Gold standard segmentation labels were manually created by 4 dentists. As a benchmark, we also evaluated MS-D networks that segmented the jaw or the teeth (i.e., binary segmentation). All segmented CBCT scans were converted to virtual 3-dimensional (3D) models. The segmentation performance of all trained MS-D networks was assessed by the Dice similarity coefficient and surface deviation. The CBCT scans segmented by the MS-D network demonstrated a large overlap with the gold standard segmentations (Dice similarity coefficient: 0.934 ± 0.019, jaw; 0.945 ± 0.021, teeth). The MS-D network–based 3D models of the jaw and the teeth showed minor surface deviations when compared with the corresponding gold standard 3D models (0.390 ± 0.093 mm, jaw; 0.204 ± 0.061 mm, teeth). The MS-D network took approximately 25 s to segment 1 CBCT scan, whereas manual segmentation took about 5 h. This study showed that multiclass segmentation of jaw and teeth was accurate and its performance was comparable to binary segmentation. The MS-D network trained for multiclass segmentation would therefore make patient-specific orthodontic treatment more feasible by strongly reducing the time required to segment multiple anatomic structures in CBCT scans.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Nevertheless, the MS-D network is not the only CNN architecture capable of performing multiclass segmentation. Several other CNN architectures have been implemented to perform multiclass segmentation of anatomies in brain (Chen et al 2018; Jafari et al 2019) and lung (Novikov et al 2018; Saood and Hatem 2020).…”

Section: Discussionmentioning

confidence: 99%

Multiclass CBCT Image Segmentation for Orthodontics with Deep Learning

et al. 2021

View full text Add to dashboard Cite

show abstract

“…This increased their dice coefficient and reduced training time, and prevented over-fitting [60]. Another study also had increased accuracy values when they added a batch normalization layer along with the residual block to segment the substantia nigra of the brain [61].…”

Section: Discussionmentioning

confidence: 99%

An automatic approach to fetal magnetic resonance image segmentation using 2D U-Net Architecture

Nithiyanantham¹

2023

Preprint

View full text Add to dashboard Cite

<p>Fetal magnetic resonance imaging is imperative to diagnosing and treating fetal disorders because it is the most effective imaging modality given its high spatial and coarse resolutions and soft-tissue contrast. Segmentation is a required step performed by radiologists to aid clinicians to treat and track disease in utero. Segmentation is followed by biometric calculations to determine the weight of the fetus for diagnosing intrauterine growth restrictions, fetal brain and cardiac abnormalities, and other fetal and congenital disorders. However, manual segmentations are time-consuming, inaccurate, and dependent on the skills of the operator. An automatic segmentation method can mitigate these drawbacks and improve maternal-fetal health by reducing wait times for treatment and improving the accuracy and standardization of segmentations. This thesis presents an automatic algorithm using the successful deep learning model U-Net, to segment the whole fetus from and MRI of the maternal abdomen with 86.70% Dice Coefficient accuracy. This is the first convolutional neural network applied for this task and outperforms other models used for similar tasks. This novel algorithm can be applied in both clinical and research settings as pre-processing pipelines to segment the whole fetus from maternal MR images.</p>

show abstract

“…First, an 8-layered (excluding the first layer) CNN model was developed using the following hyper-parameters: batch size 50, 60 epochs, learning rate 0.001 and Adam optimization parameters (betas 0.9, 0.999) [73] (Figure 2b). The weight map [74] from weighted loss function was used to counter the imbalanced dataset. Weight balancing helps to balance the data by changing the weight of training data, as the loss is computed.…”

Section: Gabor Cnn Deep Modelmentioning

confidence: 99%