Please use this identifier to cite or link to this item:
https://hdl.handle.net/10356/136988
Title: | Movement analysis and posture recognition using kinect v2 recordings | Authors: | Rajendran Karthika | Keywords: | Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence Engineering::Computer science and engineering::Software::Programming languages |
Issue Date: | 2019 | Publisher: | Nanyang Technological University | Abstract: | The prime objective of this dissertation work is to identify the postures of the psychiatrist participants based on the movement features while interacting with the schizophrenic patient. Schizophrenia is a chronic brain disorder where individuals expound reality abnormally and influence their everyday routine. Consulting a psychiatrist is the primary treatment for diagnosing schizophrenia. Behind several stages, the psychiatrist manually diagnoses the patients, which sometimes gives an insufficient result. To resolve this issue, the psychiatrist must be properly trained. As these patients are delicate, the psychiatrist who interacts with the schizophrenic patients should be cautious with their conduct, body motions, speech mode, and method of discourse. Thus, the interaction of the psychiatrist participant is recorded using Kinect V2, and the gestures of the psychiatrist participants are recognized and analyzed using data mining classification methods. The gestures of the psychiatrist participants are initially tested by interacting with the avatar, which is modeled like a schizophrenic patient. The interaction between the participants and the avatar is recorded in video format. To analyze human gestures, the gestures of the participants are recorded with the help of Kinect V2, and the features are extracted and converted into a JSON file using body frame extraction. The coordinate points are extracted from the JSON file using python code and converted to CSV files. Then the movement features are obtained by calculating joint angles, joint displacement, and necessary movement actions using MATLAB. The movement features are then labeled and implemented using data mining classification methods. By applying different classifiers, accuracy is obtained. Based on efficiency and performance, the classifier with the highest accuracy is chosen as the best classifier. | URI: | https://hdl.handle.net/10356/136988 | Schools: | School of Electrical and Electronic Engineering | Research Centres: | Research Techno Plaza | Fulltext Permission: | restricted | Fulltext Availability: | With Fulltext |
Appears in Collections: | EEE Theses |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
MSc- Dissertation.pdf Restricted Access | MSc- Dissertation | 1.42 MB | Adobe PDF | View/Open |
Page view(s)
334
Updated on Mar 27, 2024
Download(s)
10
Updated on Mar 27, 2024
Google ScholarTM
Check
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.