Dual-Stage Deeply Supervised Attention-Based Convolutional Neural Networks for Mandibular Canal Segmentation in CBCT Scans
Accurate segmentation of mandibular canals in lower jaws is important in dental implantology. Medical experts manually determine the implant position and dimensions from 3D CT images to avoid damaging the mandibular nerve inside the canal. In this paper, we propose a novel dual-stage deep learning-b...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-12-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/22/24/9877 |
| _version_ | 1797455345633722368 |
|---|---|
| author | Muhammad Usman Azka Rehman Amal Muhammad Saleem Rabeea Jawaid Shi-Sub Byon Sung-Hyun Kim Byoung-Dai Lee Min-Suk Heo Yeong-Gil Shin |
| author_facet | Muhammad Usman Azka Rehman Amal Muhammad Saleem Rabeea Jawaid Shi-Sub Byon Sung-Hyun Kim Byoung-Dai Lee Min-Suk Heo Yeong-Gil Shin |
| author_sort | Muhammad Usman |
| collection | DOAJ |
| description | Accurate segmentation of mandibular canals in lower jaws is important in dental implantology. Medical experts manually determine the implant position and dimensions from 3D CT images to avoid damaging the mandibular nerve inside the canal. In this paper, we propose a novel dual-stage deep learning-based scheme for the automatic segmentation of the mandibular canal. In particular, we first enhance the CBCT scans by employing the novel histogram-based dynamic windowing scheme, which improves the visibility of mandibular canals. After enhancement, we designed 3D deeply supervised attention UNet architecture for localizing the Volumes Of Interest (VOIs), which contain the mandibular canals (i.e., left and right canals). Finally, we employed the Multi-Scale input Residual UNet (MSiR-UNet) architecture to segment the mandibular canals using VOIs accurately. The proposed method has been rigorously evaluated on 500 and 15 CBCT scans from our dataset and from the public dataset, respectively. The results demonstrate that our technique improves the existing performance of mandibular canal segmentation to a clinically acceptable range. Moreover, it is robust against the types of CBCT scans in terms of field of view. |
| first_indexed | 2024-03-09T15:53:04Z |
| format | Article |
| id | doaj.art-f719a44c6a404bf09ffc66845780b010 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-09T15:53:04Z |
| publishDate | 2022-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-f719a44c6a404bf09ffc66845780b0102023-11-24T17:56:54ZengMDPI AGSensors1424-82202022-12-012224987710.3390/s22249877Dual-Stage Deeply Supervised Attention-Based Convolutional Neural Networks for Mandibular Canal Segmentation in CBCT ScansMuhammad Usman0Azka Rehman1Amal Muhammad Saleem2Rabeea Jawaid3Shi-Sub Byon4Sung-Hyun Kim5Byoung-Dai Lee6Min-Suk Heo7Yeong-Gil Shin8Center for Artificial Intelligence in Medicine and Imaging, HealthHub Co., Ltd., Seoul 06524, Republic of KoreaCenter for Artificial Intelligence in Medicine and Imaging, HealthHub Co., Ltd., Seoul 06524, Republic of KoreaCenter for Artificial Intelligence in Medicine and Imaging, HealthHub Co., Ltd., Seoul 06524, Republic of KoreaDivision of AI and Computer Engineering, Kyonggi University, Suwon 16227, Republic of KoreaCenter for Artificial Intelligence in Medicine and Imaging, HealthHub Co., Ltd., Seoul 06524, Republic of KoreaCenter for Artificial Intelligence in Medicine and Imaging, HealthHub Co., Ltd., Seoul 06524, Republic of KoreaDivision of AI and Computer Engineering, Kyonggi University, Suwon 16227, Republic of KoreaDepartment of Oral and Maxillofacial Radiology, School of Dentistry, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Republic of KoreaDepartment of Computer Science and Engineering, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Republic of KoreaAccurate segmentation of mandibular canals in lower jaws is important in dental implantology. Medical experts manually determine the implant position and dimensions from 3D CT images to avoid damaging the mandibular nerve inside the canal. In this paper, we propose a novel dual-stage deep learning-based scheme for the automatic segmentation of the mandibular canal. In particular, we first enhance the CBCT scans by employing the novel histogram-based dynamic windowing scheme, which improves the visibility of mandibular canals. After enhancement, we designed 3D deeply supervised attention UNet architecture for localizing the Volumes Of Interest (VOIs), which contain the mandibular canals (i.e., left and right canals). Finally, we employed the Multi-Scale input Residual UNet (MSiR-UNet) architecture to segment the mandibular canals using VOIs accurately. The proposed method has been rigorously evaluated on 500 and 15 CBCT scans from our dataset and from the public dataset, respectively. The results demonstrate that our technique improves the existing performance of mandibular canal segmentation to a clinically acceptable range. Moreover, it is robust against the types of CBCT scans in terms of field of view.https://www.mdpi.com/1424-8220/22/24/9877mandibular canal3D segmentationjaw localizationCBCT |
| spellingShingle | Muhammad Usman Azka Rehman Amal Muhammad Saleem Rabeea Jawaid Shi-Sub Byon Sung-Hyun Kim Byoung-Dai Lee Min-Suk Heo Yeong-Gil Shin Dual-Stage Deeply Supervised Attention-Based Convolutional Neural Networks for Mandibular Canal Segmentation in CBCT Scans Sensors mandibular canal 3D segmentation jaw localization CBCT |
| title | Dual-Stage Deeply Supervised Attention-Based Convolutional Neural Networks for Mandibular Canal Segmentation in CBCT Scans |
| title_full | Dual-Stage Deeply Supervised Attention-Based Convolutional Neural Networks for Mandibular Canal Segmentation in CBCT Scans |
| title_fullStr | Dual-Stage Deeply Supervised Attention-Based Convolutional Neural Networks for Mandibular Canal Segmentation in CBCT Scans |
| title_full_unstemmed | Dual-Stage Deeply Supervised Attention-Based Convolutional Neural Networks for Mandibular Canal Segmentation in CBCT Scans |
| title_short | Dual-Stage Deeply Supervised Attention-Based Convolutional Neural Networks for Mandibular Canal Segmentation in CBCT Scans |
| title_sort | dual stage deeply supervised attention based convolutional neural networks for mandibular canal segmentation in cbct scans |
| topic | mandibular canal 3D segmentation jaw localization CBCT |
| url | https://www.mdpi.com/1424-8220/22/24/9877 |
| work_keys_str_mv | AT muhammadusman dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans AT azkarehman dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans AT amalmuhammadsaleem dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans AT rabeeajawaid dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans AT shisubbyon dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans AT sunghyunkim dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans AT byoungdailee dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans AT minsukheo dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans AT yeonggilshin dualstagedeeplysupervisedattentionbasedconvolutionalneuralnetworksformandibularcanalsegmentationincbctscans |