http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Mechanical Design of Robot Head with Human-like Emotions
Nguyen Khac Toan,Le Duc Thuan,Le Bao Long,Nguyen Truong Thinh 제어로봇시스템학회 2021 제어로봇시스템학회 국제학술대회 논문집 Vol.2021 No.10
One of the fields that has detonated lately dependent on the AI platform is humanoid robots. The facial expressions of robots, then again, are getting less and less attention. The design of a robot head that communicates a wide range of human-like expressions is described in this paper. The eye-eyelid, lip-jaw, and neck were the three main parts of the robot head under investigation. Mechanical structure of human head such as number of degrees of freedom, size, ability of joints, and major muscle groups to create movements in each part are determined based on anatomy. Then, facial expressions are analyzed dependent on AUs to find out reasonable control points. The mechanical structures of each part were designed to try to replicate the movement and expression as closely as expected. Additionally, proportions, sizes, distances of the designs, such as CoP-CoP, tr-n, n-sn, and so on, must closely resemble those of a human head. The robot head in this article is designed and modeled on an Asian human head. The result of the design, the head section is made up of three mechanisms connected by a shaft that runs through the entire head. Through the experimental process, the robot head can perform 6 of the basic expressions and movements like a real human head.
Facial landmarks detection for evaluating facial paralysis using a modern active shape model
Hung Phuoc Truong,Quan Manh Le,Thinh Long Nguyen,Yong-Guk Kim 한국HCI학회 2018 한국HCI학회 학술대회 Vol.2018 No.1
Facial nerve palsy makes patients lose voluntary muscles movement typically at one side of the human face and then it makes them difficult to interact with people using natural facial expressions. Detection of such symptom and evaluation of how serious it is should be a big challenge for standard face models. We propose a new method by which facial landmarks of the palsy face can be detected and measure the degree of the symptom using Discriminative Bayesian Active Shape Model (DB-ASM). Each patient was asked to make a few facial actions, such as eye closing, E-action and O-action with his mouth. We use facial landmarks detection method in measuring displacement ratios reflecting the structure of regions of eyebrows, eyes and mouth, respectively. There are three steps for this: annotating face model, fitting facial landmarks and measuring symmetry indices on the face. The present dataset contains many videos collected from 14 patients and 6 healthy persons. Experimental result suggests that the model with DB-ASM is useful as a facial landmarks predictor for facial paralysis patients.