scholarly journals Evaluation of multi-task learning in deep learning-based positioning classification of mandibular third molars

2022 ◽  
Vol 12 (1) ◽  
Author(s):  
Shintaro Sukegawa ◽  
Tamamo Matsuyama ◽  
Futa Tanaka ◽  
Takeshi Hara ◽  
Kazumasa Yoshii ◽  
...  
Keyword(s):  
Deep Learning ◽  
Performance Metrics ◽  
Maxillofacial Surgery ◽  
Effect Sizes ◽  
Medium Effect ◽  
Third Molars ◽  
Single Task ◽  
Task Learning ◽  
Mandibular Third Molars

AbstractPell and Gregory, and Winter’s classifications are frequently implemented to classify the mandibular third molars and are crucial for safe tooth extraction. This study aimed to evaluate the classification accuracy of convolutional neural network (CNN) deep learning models using cropped panoramic radiographs based on these classifications. We compared the diagnostic accuracy of single-task and multi-task learning after labeling 1330 images of mandibular third molars from digital radiographs taken at the Department of Oral and Maxillofacial Surgery at a general hospital (2014–2021). The mandibular third molar classifications were analyzed using a VGG 16 model of a CNN. We statistically evaluated performance metrics [accuracy, precision, recall, F1 score, and area under the curve (AUC)] for each prediction. We found that single-task learning was superior to multi-task learning (all p < 0.05) for all metrics, with large effect sizes and low p-values. Recall and F1 scores for position classification showed medium effect sizes in single and multi-task learning. To our knowledge, this is the first deep learning study to examine single-task and multi-task learning for the classification of mandibular third molars. Our results demonstrated the efficacy of implementing Pell and Gregory, and Winter’s classifications for specific respective tasks.

scholarly journals A Deep Learning-based Positioning Classification of the Mandibular Third Molars: Is Multi-task Deep Learning Useful?

2021 ◽  
Author(s):  
Shintaro Sukegawa ◽  
Tamamo Matsuyama ◽  
Futa Tanaka ◽  
Takeshi Hara ◽  
Kazumasa Yoshii ◽  
...  
Keyword(s):  
Deep Learning ◽  
Performance Metrics ◽  
Maxillofacial Surgery ◽  
Effect Sizes ◽  
Medium Effect ◽  
Third Molars ◽  
Single Task ◽  
Task Learning ◽  
Mandibular Third Molars

Abstract Pell and Gregory, and Winter’s classifications are frequently implemented to classify the mandibular third molars and are crucial for safe tooth extraction. This study aimed to evaluate the classification accuracy of convolutional neural network (CNN) deep learning models using cropped panoramic radiographs based on these classifications. We compared the diagnostic accuracy of single-task and multi-task learning after labeling 1,330 images of mandibular third molars from digital radiographs taken at the Department of Oral and Maxillofacial Surgery at a general hospital (2014-2021). The mandibular third molar classifications were analyzed using a VGG 16 model of a CNN. We statistically evaluated performance metrics (accuracy, precision, recall, F1 score, area under the curve [AUC]) for each prediction. We found that single-task learning was superior to multi-task learning (all p<0.05) for all metrics, with large effect sizes and low p-values. Recall and F1 scores for position classification showed medium effect sizes in single and multi-task learning. To our knowledge, this deep learning study is the first to examine single-task and multi-task learning for the classification of mandibular third molars. Our results demonstrated the efficacy of implementing Pell and Gregory, and Winter’s classifications for specific respective tasks.

Abstract 279: Multi-task Learning Improves Model Performance in Predicting Rare Catastrophic Events in Healthcare Claims Dataset

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_4) ◽  
Author(s):  
ChienYu Chi ◽  
Yen-Pin Chen ◽  
Adrian Winkler ◽  
Kuan-Chun Fu ◽  
Fie Xu ◽  
...  
Keyword(s):  
Cardiac Arrest ◽  
Deep Learning ◽  
Short Term Memory ◽  
Learning System ◽  
Research Database ◽  
Learning Models ◽  
Catastrophic Events ◽  
Single Task ◽  
Task Learning ◽  
Hospital Cardiac Arrest

Introduction: Predicting rare catastrophic events is challenging due to lack of targets. Here we employed a multi-task learning method and demonstrated that substantial gains in accuracy and generalizability was achieved by sharing representations between related tasks Methods: Starting from Taiwan National Health Insurance Research Database, we selected adult people (>20 year) experienced in-hospital cardiac arrest but not out-of-hospital cardiac arrest during 8 years (2003-2010), and built a dataset using de-identified claims of Emergency Department (ED) and hospitalization. Final dataset had 169,287 patients, randomly split into 3 sections, train 70%, validation 15%, and test 15%.Two outcomes, 30-day readmission and 30-day mortality are chosen. We constructed the deep learning system in two steps. We first used a taxonomy mapping system Text2Node to generate a distributed representation for each concept. We then applied a multilevel hierarchical model based on long short-term memory (LSTM) architecture. Multi-task models used gradient similarity to prioritize the desired task over auxiliary tasks. Single-task models were trained for each desired task. All models share the same architecture and are trained with the same input data Results: Each model was optimized to maximize AUROC on the validation set with the final metrics calculated on the held-out test set. We demonstrated multi-task deep learning models outperform single task deep learning models on both tasks. While readmission had roughly 30% positives and showed miniscule improvements, the mortality task saw more improvement between models. We hypothesize that this is a result of the data imbalance, mortality occurred roughly 5% positive; the auxiliary tasks help the model interpret the data and generalize better. Conclusion: Multi-task deep learning models outperform single task deep learning models in predicting 30-day readmission and mortality in in-hospital cardiac arrest patients.

scholarly journals Guitar Chord Sensing and Recognition Using Multi-Task Learning and Physical Data Augmentation with Robotics

Sensors ◽  
2020 ◽  
Vol 20 (21) ◽  
pp. 6077
Author(s):  
Gerelmaa Byambatsogt ◽  
Lodoiravsal Choimaa ◽  
Gou Koutaki
Keyword(s):  
Deep Learning ◽  
Data Augmentation ◽  
Recognition Task ◽  
Learning Model ◽  
Training Data ◽  
Single Task ◽  
Task Learning ◽  
Physical Data ◽  
Music Information ◽  
Chord Recognition

In recent years, many researchers have shown increasing interest in music information retrieval (MIR) applications, with automatic chord recognition being one of the popular tasks. Many studies have achieved/demonstrated considerable improvement using deep learning based models in automatic chord recognition problems. However, most of the existing models have focused on simple chord recognition, which classifies the root note with the major, minor, and seventh chords. Furthermore, in learning-based recognition, it is critical to collect high-quality and large amounts of training data to achieve the desired performance. In this paper, we present a multi-task learning (MTL) model for a guitar chord recognition task, where the model is trained using a relatively large-vocabulary guitar chord dataset. To solve data scarcity issues, a physical data augmentation method that directly records the chord dataset from a robotic performer is employed. Deep learning based MTL is proposed to improve the performance of automatic chord recognition with the proposed physical data augmentation dataset. The proposed MTL model is compared with four baseline models and its corresponding single-task learning model using two types of datasets, including a human dataset and a human combined with the augmented dataset. The proposed methods outperform the baseline models, and the results show that most scores of the proposed multi-task learning model are better than those of the corresponding single-task learning model. The experimental results demonstrate that physical data augmentation is an effective method for increasing the dataset size for guitar chord recognition tasks.

scholarly journals Benchmarking feature quality assurance strategies for non-targeted metabolomics

2021 ◽  
Author(s):  
Yasin El Abiead ◽  
Maximilian Milford ◽  
Harald Schoeny ◽  
Mate Rusz ◽  
Reza M Salek ◽  
...  
Keyword(s):  
Quality Control ◽  
Quality Assurance ◽  
Liquid Chromatography ◽  
Deep Learning ◽  
Performance Metrics ◽  
Targeted Metabolomics ◽  
Mass Spec ◽  
Chromatographic Peaks ◽  
Resolution Mass

Automated data pre-processing (DPP) forms the basis of any liquid chromatography-high resolution mass spec-trometry-driven non-targeted metabolomics experiment. However, current strategies for quality control of this im-portant step have rarely been investigated or even discussed. We exemplified how reliable benchmark peak lists could be generated for eleven publicly available datasets acquired across different instrumental platforms. Moreover, we demonstrated how these benchmarks can be utilized to derive performance metrics for DPP and tested whether these metrics can be generalized for entire datasets. Relying on this principle, we cross-validated different strategies for quality assurance of DPP, including manual parameter adjustment, variance of replicate injection-based metrics, unsupervised clustering performance, automated parameter optimization, and deep learning-based classification of chromatographic peaks. Overall, we want to highlight the importance of assessing DPP performance on a regular basis.

scholarly journals Comparison of Complications of Extraction among Partially Impacted Mandibular Third Molars with or without a Buccal Flap

2021 ◽  
pp. 12-19
Author(s):  
M. Ahmed Khan ◽  
Tahera Ayub ◽  
Bibi Gulsama ◽  
Azizullah Muhammad Nawaz Qureshi ◽  
Aosaf Anwar Memon ◽  
...  
Keyword(s):  
Operative Time ◽  
Maxillofacial Surgery ◽  
Operating Time ◽  
Periodontal Pocket ◽  
Third Molars ◽  
Significant Difference ◽  
Impacted Mandibular Third Molars ◽  
Buccal Flap ◽  
Mandibular Third Molars

Objective: To compare the complications of extraction of partially impacted mandibular third molars with or without a buccal flap. Materials And Methods: A comparative cohort study was performed at Department of Oral & Maxillofacial Surgery, Institute of Dentistry, Liaquat University Hospital, Hyderabad from September 2020 to March 2021. Sixty-two patients of either gender, having age 15-50 years and recommended for extraction of partially impacted mandibular third molars were selected by non-probability consecutive sampling technique and distributed into flapless group (31 patients) and buccal flap group (31 patients). Patients were treated with standard procedures of flapless and buccal flap, operating time was noted and follow up was done at 1st day, 2nd day post-operatively for pain, swelling, trismus, whereas periodontal pocket distal to second molar was measured at 1 month and 3 months follow up interval. Results: In flapless and buccal flap group male patients were 17 (54.8%) and 18 (58.1%) and female patients were 14 (45.2%) and 13 (41.9%) respectively with mean age of 27.4 ± 9.6 and 26.7 ± 8.4 years. Statistically significant difference was obtained in flapless and buccal flap groups in terms of operative time, pain score, swelling score, pocket depth and trismus. Conclusion: Flapless technique is more effective in conditions of operative time and post-operative complications. So, flapless technique can be used frequently for elimination of incompletely impacted mandibular third molars.

scholarly journals Automated Prediction of Extraction Difficulty and Inferior Alveolar Nerve Injury for Mandibular Third Molar Using a Deep Neural Network

2022 ◽  
Vol 12 (1) ◽  
pp. 475
Author(s):  
Junseok Lee ◽  
Jumi Park ◽  
Seong Yong Moon ◽  
Kyoobin Lee
Keyword(s):  
Maxillofacial Surgery ◽  
Inferior Alveolar Nerve ◽  
Third Molar ◽  
Oral And Maxillofacial Surgery ◽  
Third Molars ◽  
Radiographic Image ◽  
Common Procedure ◽  
Radiographic Images ◽  
Mandibular Third Molar ◽  
Mandibular Third Molars

Extraction of mandibular third molars is a common procedure in oral and maxillofacial surgery. There are studies that simultaneously predict the extraction difficulty of mandibular third molar and the complications that may occur. Thus, we propose a method of automatically detecting mandibular third molars in the panoramic radiographic images and predicting the extraction difficulty and likelihood of inferior alveolar nerve (IAN) injury. Our dataset consists of 4903 panoramic radiographic images acquired from various dental hospitals. Seven dentists annotated detection and classification labels. The detection model determines the mandibular third molar in the panoramic radiographic image. The region of interest (ROI) includes the detected mandibular third molar, adjacent teeth, and IAN, which is cropped in the panoramic radiographic image. The classification models use ROI as input to predict the extraction difficulty and likelihood of IAN injury. The achieved detection performance was 99.0% mAP over the intersection of union (IOU) 0.5. In addition, we achieved an 83.5% accuracy for the prediction of extraction difficulty and an 81.1% accuracy for the prediction of the likelihood of IAN injury. We demonstrated that a deep learning method can support the diagnosis for extracting the mandibular third molar.

scholarly journals Pattern of Angulations and Common Indications for Extraction of Impacted Mandibular Third Molar at KMU-Institute of Dental Sciences, Kohat

2019 ◽  
Vol 8 (2) ◽  
pp. 79-83
Author(s):  
Tariq Sardar ◽  
Gulrukh Sheikh ◽  
Saddique Aslam ◽  
Numan Muhammad Khan ◽  
Javed Akhtar Rana
Keyword(s):  
Maxillofacial Surgery ◽  
Third Molar ◽  
Age Group ◽  
Third Molars ◽  
Impacted Tooth ◽  
Female Ratio ◽  
Impacted Mandibular Third Molar ◽  
Mandibular Third Molar ◽  
Impacted Mandibular Third Molars ◽  
Mandibular Third Molars

Background: The extraction of an impacted mandibular third molar (MTM), with associated pathologies or clinical manifestations is an important and one of the most frequent decisions in dentistry. The angle formed by the longitudinal axis of second and third molar is used to determine angulation of impacted MTM. The aim of this study was to identify the pattern of angulations of impacted mandibular third molar and common indications for extraction associated with these angulations.Material and Methods: This descriptive cross-sectional study was carried out at Department of Oral & Maxillofacial Surgery, Khyber Medical University Institute of Dental Sciences, Kohat, Khyber Pakhtunkhwa (KP) from November 2017 to July 2018. A total of 349 patients presenting with impacted mandibular third molars were included in this study. Name, age, gender, address, the angulation of the impacted tooth and the indication for extraction of the impacted tooth were recorded. Data comprising of qualitative and quantitative variables were analyzed using SPSS version 17.Results: Out of 349 patients, 206 were male and 143 females, with the male to female ratio of 1.4:1. The age range of the patients was from 18 years to 60 years with a mean age of 26 ± 6 years. The most common age group with impacted third molar was ≤ 25 years followed by 26 to 30 years’ age group. The most common angulation was mesioangular followed by vertical, horizontal and distoangular impacted mandibular third molar. Pericoronitis was the most common indication for extraction in all angulations except horizontal impaction where root resorption of the second molar was more common.Conclusion: Mesioangular is the most common angulation in impacted mandibular third molars. Pericoronitis is the main indication for all angulations of impacted mandibular third molars except horizontal angulation, occurring mostly in the third decade of life.

scholarly journals Artificial Intelligence Model to Detect Real Contact Relationship between Mandibular Third Molars and Inferior Alveolar Nerve Based on Panoramic Radiographs

Diagnostics ◽  
2021 ◽  
Vol 11 (9) ◽  
pp. 1664
Author(s):  
Tianer Zhu ◽  
Daqian Chen ◽  
Fuli Wu ◽  
Fudong Zhu ◽  
Haihua Zhu
Keyword(s):  
Artificial Intelligence ◽  
Deep Learning ◽  
Inferior Alveolar Nerve ◽  
Third Molars ◽  
Panoramic Radiographs ◽  
The Real ◽  
Detection Model ◽  
Cooperative Approach ◽  
Real Contact ◽  
Mandibular Third Molars

This study aimed to develop a novel detection model for automatically assessing the real contact relationship between mandibular third molars (MM3s) and the inferior alveolar nerve (IAN) based on panoramic radiographs processed with deep learning networks, minimizing pseudo-contact interference and reducing the frequency of cone beam computed tomography (CBCT) use. A deep-learning network approach based on YOLOv4, named as MM3-IANnet, was applied to oral panoramic radiographs for the first time. The relationship between MM3s and the IAN in CBCT was considered the real contact relationship. Accuracy metrics were calculated to evaluate and compare the performance of the MM3–IANnet, dentists and a cooperative approach with dentists and the MM3–IANnet. Our results showed that in comparison with detection by dentists (AP = 76.45%) or the MM3–IANnet (AP = 83.02%), the cooperative dentist–MM3–IANnet approach yielded the highest average precision (AP = 88.06%). In conclusion, the MM3-IANnet detection model is an encouraging artificial intelligence approach that might assist dentists in detecting the real contact relationship between MM3s and IANs based on panoramic radiographs.

scholarly journals Inferior alveolar nerve injury assessment after surgical removal of mandibular third molar.

2020 ◽  
Vol 27 (03) ◽  
pp. 530-534
Author(s):  
Abdul Wahid Bhangwar ◽  
Muhammad Irfan Khan ◽  
Hira Fatima ◽  
Salman Shams
Keyword(s):  
Nerve Injury ◽  
Local Anesthesia ◽  
Maxillofacial Surgery ◽  
Inferior Alveolar Nerve ◽  
Third Molar ◽  
Surgical Removal ◽  
Third Molars ◽  
Mandibular Third Molar ◽  
Mandibular Third Molars ◽  
Inferior Alveolar Nerve Injury

To assess the nerve injury (inferior alveolar nerve) after surgical removal of mandibular third molars under local anesthesia. Study Design: Observational study. Setting: Oral & Maxillofacial Surgery Department LUMHS Jamshoro/Hyderabad. Period: From 11th November 2015 to 10th May 2016. Material & Methods: This study consisted of one hundred patients. Inclusion criteria’s were patients with impacted mandibular third molar, patient’s age from 18 to 45years and irrespective of gender. Exclusion criteria were patients younger than 18yrs of age of above 45 years, patients having neurological disorders, medically compromised patients, patients receiving radiotherapy or chemotherapy, patients with known allergy to local anesthesia, patients having pathology due to mandibular third molars, patients radiographicaly root is very near to inferior dental canal. Results: Out of 100 patients incorporated in this research 66 were male (66%) and 34 female (34%). The mean age was 29+3.20 years. Common indication of extraction were recurrent pericoronitis  52(52%) cases followed by deep caries/ pulpitis in 28(28%)  cases, orthodontic reason in 11(11%) cases and caries to adjacent tooth in 9(9%) cases. Third molar impaction according to winter’s classification were Mesioangular in 54(54%) cases followed by Horizontal in 26(26%) cases and Vertical in 11(11%). Radiographic showed Narrowing of root in 21% cases and narrowing of inferior dental (ID) canal 20% cases, followed by diversion of ID canal in 16 % cases, deflection of root 14 % cases and darkening of root in 11% cases. After surgical removal of mandibular third molar, the inferior alveolar nerve injury was observed in 6(6%) cases. Conclusion: We conclude that inferior alveolar nerve paresthesia occurs in 6% after surgical removal of mandibular third molars.

Sign in / Sign up

Export Citation Format

Share Document