A fully deep learning model for the automatic identification of cephalometric landmarks

Kim, Young Hyun,Lee, Chena,Ha, Eun-Gyu,Choi, Yoon Jeong,Han, Sang-Sun Korean Academy of Oral and Maxillofacial Radiology 2021 Imaging Science in Dentistry Vol.51 No.3

Purpose: This study aimed to propose a fully automatic landmark identification model based on a deep learning algorithm using real clinical data and to verify its accuracy considering inter-examiner variability. Materials and Methods: In total, 950 lateral cephalometric images from Yonsei Dental Hospital were used. Two calibrated examiners manually identified the 13 most important landmarks to set as references. The proposed deep learning model has a 2-step structure-a region of interest machine and a detection machine-each consisting of 8 convolution layers, 5 pooling layers, and 2 fully connected layers. The distance errors of detection between 2 examiners were used as a clinically acceptable range for performance evaluation. Results: The 13 landmarks were automatically detected using the proposed model. Inter-examiner agreement for all landmarks indicated excellent reliability based on the 95% confidence interval. The average clinically acceptable range for all 13 landmarks was 1.24 mm. The mean radial error between the reference values assigned by 1 expert and the proposed model was 1.84 mm, exhibiting a successful detection rate of 36.1%. The A-point, the incisal tip of the maxillary and mandibular incisors, and ANS showed lower mean radial error than the calibrated expert variability. Conclusion: This experiment demonstrated that the proposed deep learning model can perform fully automatic identification of cephalometric landmarks and achieve better results than examiners for some landmarks. It is meaningful to consider between-examiner variability for clinical applicability when evaluating the performance of deep learning methods in cephalometric landmark identification.

이동로봇의 저비용 위치추정을 위한 효율적인 인공표식 배치기법

김지웅,정우진,Kim, Jiwoong,Chung, Woojin 한국전기전자학회 2013 전기전자학회논문지 Vol.17 No.4

인공표식은 이동로봇의 위치추정에서 불확실성을 감소시키기 위해 널리 사용되어 왔다. 또한, 사용되는 인공표식의 수가 증가함에 따라 위치추정 비용이 증가하기 때문에, 인공표식의 효율적 배치를 위한 연구는 핵심적인 이슈중 하나로서 여겨져 왔다. 따라서 본 논문은 운동모델과 센서모델의 불확실성 특성을 고려함으로써 인공표식들을 효율적으로 배치하기 위한 방법을 제안한다. 운동모델과 센서모델은 서로 다른 불확실성 분포를 가지기 때문에, 최종 불확실성은 두 가지 불확실성의 효율적 조합을 통해 크게 감소될 수 있다. 제안한 기법의 유용성은 시뮬레이션 결과에 의해 입증된다. Artificial landmarks have been widely used for reducing the uncertainty in localization of a mobile robot. In addition, research for efficient placement of artificial landmarks has been considered as one of the fundamental issues since the cost of localization is increased with the number of used landmarks. Therefore, this paper proposes a method in which landmarks are efficiently placed by considering the uncertainty characteristics of the motion model and the sensor model. Because two models have different uncertainty distributions, the final uncertainty can be considerably reduced through their efficient combination. The usefulness of the proposed method is demonstrated by simulation results.

Implementation of real-time positioning system using extended Kalman filter and artificial landmark on ceiling

Rusdinar, Angga,Kim, Jung-Min,Lee, Jun-Ha,Kim, Sung-Shin 대한기계학회 2012 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.26 No.3

Most localization algorithms use a range sensor or vision in a horizontal view, which usually imparts some disruption from a dynamic or static obstacle. By using landmarks on ceiling which the vehicle position were vertically measured, the disruption from horizontal view was reduced. We propose an indoor localization and navigation system based on an extended Kalman filter (EKF) and real-time vision system. A single upward facing digital camera was mounted on an autonomous vehicle as a vision sensor to recognize the landmarks. The landmarks consisted of multiple circles that were arranged in a defined pattern. Information on a landmark's direction and its identity as a reference for an autonomous vehicle was produced by the circular arrangements. The pattern of the circles was detected using a robust image processing algorithm. To reduce the noise that came from uneven light, the process of noise reduction was separated into several regions of interest. The accumulative error caused by odometry sensors (i.e., encoders and a gyro) and the vehicle's position were calculated and estimated, respectively, using the EKF algorithm. Both algorithms were tested on a vehicle in a real environment. The image processing method could precisely recognize the landmarks, and the EKF algorithm could accurately estimate the vehicle's position. The experimental results confirmed that the proposed approaches are implementable.

Hand Landmark를 이용한 한글 지문자 인식

김진영,강의성 한국차세대컴퓨팅학회 2022 한국차세대컴퓨팅학회 논문지 Vol.18 No.1

Sign Language are languages that use a visual means of communicating using gestures and facial expression. Finger spelling uses hand movements and shapes to spell out the proper nouns or words, which are difficult to be represented. In this study, Korean finger spelling recognition system is proposed. Landmarks are extracted from hand images by utilizing MediaPipe Hands API. Also SVM(Support Vector Machine) of scikit-learn is used to train 31 Korean finger spellings. Finger spelling such as a double consonant ‘ㄲ’ is represented by moving ‘ㄱ’ in the right direction. The proposed method recognized ‘ㄲ’ by detecting the movement of the bounding box centroid for finger spelling. Most of the previous methods are not able to recognize all the finger spellings for Korean sign language, but the proposed technique can do all 36 Korean finger spellings. Also, it is expected that our methods are utilized in finger spelling learning or sign language translation for hearing-impaired persons. 수어는 음성이 아닌 시각적 제스처(Gesture)를 매체로 의미를 전달한다. 시각적인 제스처를 이용하여 의미를 주고받기 어려운 단어나 고유명사를 전달하기 위해서는 손과 손가락의 모양으로 한글의 자음과 모음을 표현한 지문자가 주로 사용된다. 본 연구에서는 MediaPipe Hands API를 이용하여 손 이미지로부터 Landmark를 추출하고 scikit-learn SVM(Support Vector Machine)을 이용하여 31개의 한글 지문자에 대한 학습 모델을 생성한다. ‘ㄲ’과 같은 지문자는 ‘ㄱ’에 대한 지문자가 오른쪽으로 이동하는 제스처로 표현되는데, 제안한 방법에서는 ‘ㄱ’ 무게 중심이 이동하는 것을 검출하여 ‘ㄲ’을 인식할 수 있다. 기존의 연구에서는 특정 수어 문장 또는 특정 지문자를 제한적으로 인식하는 연구가 주를 이루고 있지만, 제안한 방법에서는 한글 지문자를 구성하는 36개의 지문자를 인식할 수 있기 때문에 청각장애인들의 지문자 학습, 수어 번역 등을 위해 유용하게 이용될 수 있을 것으로 기대된다.

레이저 스캐너와 마커센서를 사용한 무인운반차의 위치추정시스템

허성우(Sung-Woo Heo),박태형(Tae Hyoung Park) 제어로봇시스템학회 2017 제어·로봇·시스템학회 논문지 Vol.23 No.10

This study develops a localization system using marker sensors and a laser scanner for autonomous guided vehicles (AGVs). To calibrate the relative position data of robots, which are obtained from encoders and gyros, the system uses a laser scanner and magnetic markers. The existing position-tracking method distinguish between rail and non-rail. This study proposes a method that enhances the position accuracy of robots using marker sensors and a laser scanner without distinguishing between rail and non-rail. First, using a laser scanner, the robot positions are obtained by measuring the distance between the reflector and the robot. The magnetic marker sensors provide the absolute position by detecting magnetic landmarks. Using a Kalman filter to combine the position data retrieved from the two sensors with the position data from the encoders and gyro sensors, we propose an effective localization system that integrates rails and non-rails. The experiment was repeated at three places using an actual AGV that was developed for this study. The experimental results demonstrate that the proposed system can effectively track the positions of AGVs.

Vision-Based Indoor Localization Using Artificial Landmarks and Natural Features on the Ceiling with Optical Flow and a Kalman Filter

Angga Rusdinar,Sungshin Kim 한국지능시스템학회 2013 INTERNATIONAL JOURNAL of FUZZY LOGIC and INTELLIGE Vol.13 No.2

This paper proposes a vision-based indoor localization method for autonomous vehicles. A single upward-facing digital camera was mounted on an autonomous vehicle and used as a vision sensor to identify artificial landmarks and any natural corner features. An interest point detector was used to find the natural features. Using an optical flow detection algorithm, information related to the direction and vehicle translation was defined. This information was used to track the vehicle movements. Random noise related to uneven light disrupted the calculation of the vehicle translation. Thus, to estimate the vehicle translation, a Kalman filter was used to calculate the vehicle position. These algorithms were tested on a vehicle in a real environment. The image processing method could recognize the landmarks precisely, while the Kalman filter algorithm could estimate the vehicle’s position accurately. The experimental results confirmed that the proposed approaches can be implemented in practical situations.

Vision-Based Indoor Localization Using Artificial Landmarks and Natural Features on the Ceiling with Optical Flow and a Kalman Filter

Rusdinar, Angga,Kim, Sungshin Korean Institute of Intelligent Systems 2013 INTERNATIONAL JOURNAL of FUZZY LOGIC and INTELLIGE Vol.13 No.2

This paper proposes a vision-based indoor localization method for autonomous vehicles. A single upward-facing digital camera was mounted on an autonomous vehicle and used as a vision sensor to identify artificial landmarks and any natural corner features. An interest point detector was used to find the natural features. Using an optical flow detection algorithm, information related to the direction and vehicle translation was defined. This information was used to track the vehicle movements. Random noise related to uneven light disrupted the calculation of the vehicle translation. Thus, to estimate the vehicle translation, a Kalman filter was used to calculate the vehicle position. These algorithms were tested on a vehicle in a real environment. The image processing method could recognize the landmarks precisely, while the Kalman filter algorithm could estimate the vehicle's position accurately. The experimental results confirmed that the proposed approaches can be implemented in practical situations.

천정 인공 랜드마크에 의한 이동로봇의 정밀 위치추정에 관한 연구

채문석(Moon-Seok Chae),양태규(Tae-Kyu Yang) 한국정보기술학회 2011 한국정보기술학회논문지 Vol.9 No.8

The technique of simultaneous localization is the most important research topic in mobile robotics. In this study we presented a precise localization algorithm for mobile robot based on artificial landmarks in the ceiling. This technique is estimated location of landmarks on the ceiling and generated the global ceiling map for landmarks, estimated the location of a mobile robot based on the ceiling map. This localization algorithm can be removed incorrectly recognized landmark using a histogram and the measurement error using Kalman filter. Experimental environment was constructed to evaluate the performance of proposed precise localization for mobile robot in a real indoor, the location of the mobile robot is recognized by using StarGazer which uses artificial landmarks attached on the ceiling. Experiments have demonstrated the feasibility of the proposed precise localization algorithm.

사각지대를 고려한 이동로봇의 인공표식기반 위치추정시스템

허동혁,박태형 한국로봇학회 2011 로봇학회 논문지 Vol.6 No.2

This paper propose a localization system of indoor mobile robots. The localization system includes camera and artificial landmarks for global positioning, and encoders and gyro sensors for local positioning. The Kalman filter is applied to take into account the stochastic errors of all sensors. Also we develop a dead reckoning system to estimate the global position when the robot moves the blind spots where it cannot see artificial landmarks, The learning engine using modular networks is designed to improve the performance of the dead reckoning system. Experimental results are then presented to verify the usefulness of the proposed localization system.

A Study on the Screening of Children at Risk for Developmental Disabilities Using Facial Landmarks Derived From a Mobile-Based Application

Sang Ho Hwang,Yeonsoo Yu,Jichul Kim,Taeyeop Lee,Yu Rang Park,Hyo-Won Kim 대한신경정신의학회 2024 PSYCHIATRY INVESTIGATION Vol.21 No.5

Objective : Early detection and intervention of developmental disabilities (DDs) are critical to improving the long-term outcomes of af-flicted children. In this study, our objective was to utilize facial landmark features from mobile application to distinguish between children with DDs and typically developing (TD) children. Methods : The present study recruited 89 children, including 33 diagnosed with DD, and 56 TD children. The aim was to examine the effectiveness of a deep learning classification model using facial video collected from children through mobile-based application. The study participants underwent comprehensive developmental assessments, which included the child completion of the Korean Psychoedu-cational Profile-Revised and caregiver completing the Korean versions of Vineland Adaptive Behavior Scale, Korean version of the Child-hood Autism Rating Scale, Social Responsiveness Scale, and Child Behavior Checklist. We extracted facial landmarks from recorded vid-eos using mobile application and performed DDs classification using long short-term memory with stratified 5-fold cross-validation. Results : The classification model shows an average accuracy of 0.88 (range: 0.78–1.00), an average precision of 0.91 (range: 0.75–1.00), and an average F1-score of 0.80 (range: 0.60–1.00). Upon interpreting prediction results using SHapley Additive exPlanations (SHAP), we verified that the most crucial variable was the nodding head angle variable, with a median SHAP score of 2.6. All the top 10 contributing variables exhibited significant differences in distribution between children with DD and TD (p<0.05). Conclusion : The results of this study provide evidence that facial landmarks, utilizing readily available mobile-based video data, can be used to detect DD at an early stage.