"CSG6224 (SPECIAL TOPIC)– PROJECT PROPOSAL SEMESTER:2NAME OF STUDENT: STUDENT NUMBER:PROJECT TITLE: PREDICTION OF OBSTRUCTIVE SLEEP APNEA FROM CRANIO FACIAL IMAGESSUPERVISOR:Aim of the project To explore the possibility of using craniofacial photographs for the prediction of Obstructivesleep apnea.Objectives of the project?To understand the meaning, causes and symptoms of obstructive sleep apnea and theirrelationship with facial structure.? To develop algorithms to extract craniofacial features to predict obstructive sleep apnea.? To develop feature selection and classification algorithms for the prediction of obstructive sleepapnea.Background of the studyIn the present era, obstructive sleep apnea disorder is continuing increasing in the patient. It is aserious sleep disorder which arises due to breathing repeatedly stop and start during the sleep. There arevarious kids of the sleep apnea but the most common is obstructive apnea.This kind of condition may arise and result of the craniofacial structure and the specific facialfeatures such as face width, eye width etc. While this kind of disorder generated in the individual thentheir face structure may different from the normal person.Obstructive sleep Apnea is common disease which is associated with sleep fragmentation andoxygen desaturation. Diagnosis of OSA is quite difficult because for giving correct treatment and toexamining the level of disease specialist have to monitor the patients overnight. In addition, this processis quite costly and resource limited. That is why majority of persons become unable to get timelytreatment for this medical issue. Review of literaturePhotographic analysis techniques support the medical professionals in detail quantitativeassessment of craniofacial morphology (Gay, Mills and Airasian,2011). It is the great techniques whichcapture risk factors related to OSA. Author has stated that relationship between surface facialdimensions, upper airways structures and craniofacial morphology are effective for prediction ofOSAHS. It is method that supports in measuring ethic difference in index and comparing surface facialdimensions as well.There is no significant difference in face's vertical dimensions such as upper face cut, lower factcut (Žikic, B., 2016). It has been stated that persons those who are suffering with OSA have great midand lower face widths. This part of the body appeared as wider and flatter as compare to other part offace (VanPatten and Williams, 2014) has argued that volume of mid face region is larger in OSAsubjects as compare to normal persons. In respect to OSA length of mandible is shorter.Obstructive sleep apnea is the common and serious disorder in which breathing can stop duringthe 10 seconds while human sleep. This kind of disorder arise due to decrease oxygen in the blood andcan briefly awaken sleepers throughout the night. ISO 14001As per the research it has been founded that obstructive sleep apnea is arisen due to the excessweight and obesity which is associated with the soft tissues can cause the airway to become blocked.Thus, it can be said that this kind of disorder arisen among the individual due to ten breathing.Individual with the OSA are not aware of the difficulty breathing. It is often recognised as a problem byother who observe the individual at the period of interaction.Proposed methodologyData Collection: There are three datasets in this project: ? Training? validation ? testing. The machine learning-based classification algorithms will be trained and validated using the trainingand validation sets. The performance of the developed algorithms will be evaluated on the testing setconstituting samples not used in the training phase.Facial Surface Features Automatic Extraction: Using Python, a 3D surface photographs will be captured. Data will be rendered in a photorealistic model for visual check. The face area will be detected automatically using a fast and accurateface detection algorithm developed by Viola and Jones and used by the CI in his biometric research.Facial features includes length of the maxilla, mandible and chin, the circumference of theneck, and the relative shape ratios (RSRs) of some surface features (e.g. length of maxilla with respectto the mandible and that of maxilla and mandible compared to the forehead and neck).Feature Selection: Features extracted will be minimised by only selecting few of them that have the mostdiscriminating attributes. Cluster classification technique will be adopted for the reduction of extraneousfeatures from the outer layers of each cluster by constantly monitoring the attributes.ISO 14001At first, appropriate noise filtering and condensation algorithms will be developed forprototype selection, then an efficient prototype construction method will be used to find the newattributes that can represent the whole data more compactly. Feature Classification and OSA Prediction:A multi-level classification technique will be developed using Machine Learning algorithmsincluding to classify the extracted features. In this method, each training sample belongs to one of thetwo di ?erent classes (positive or negative) and the goal is to construct a function which, given a newsample, will correctly predict the class to which it belongs. Different machine learning algorithmsincluding a boosting technique and deep learning will be used and their performance will be compared.Performance Evaluation: The performance of the developed algorithms will be evaluated on thetesting set comprising OSA and non-OSA subjects.Software Development: Algorithms will be developed and tested using Python.Expected outcomes To develop algorithms to extract facial features for the prediction of obstructive sleep apnea andpublish the outcome as a scholarly journal article.Proposed budgetAs all the hardware and software are either available or going to be purchased by supervisor from hisresearch funding, there will not be additional funding for this study.Legal, ethical or social considerationsAn application for human research ethics will be made before collecting the data. ISO 14001Schedule of study Wk. Wk Wk Wk. Wk. Wk. Wk. Wk. Wk.Activity 1 2 34 5 6 7 8 9 Design Research Proposal Design Literature aims and objectivesliterature reviewFocuses on the research methodology Design research proposalData collection Drafting Findings Analyzing data and surface featureextraction feature selection andclassificationSoftware developmentDesign the draftSubmissionISO 14001ReferencesCreswell, J. W., (2013). Research design: Qualitative, quantitative, and mixed methods approaches.Sage publications.VanPatten, B. and Williams, J., (2014). Theories in second language acquisition: An introduction.Routledge. Žikic, B., (2016). Qualitative Field Research in Anthropology: An Overview of Basic ResearchMethodology. Issues in Ethnology and Anthropology. 2(2). pp.123-135.Flick, U., (2015). Introducing research methodology: A beginner's guide to doinga research project. Sage.Gast, D.L. and Ledford, J.R., (2014). Single case research methodology:Applications in special education and behavioral sciences. Routledge.Gay, L. R., Mills, G. E. and Airasian, P. W., 2011. Educational research:Competencies for analysis and applications. Pearson Higher Ed.Hunleth, J., (2011). Beyond on or with: Questioning power dynamics andknowledge production in ‘child-oriented research methodology. Childhood,18(1),pp.81-93Panneerselvam, R., 2014. Research methodology. PHI Learning Pvt. Ltd.Islam, SMS et al. (2012) Three dimensional imaging based diagnosis forObstructive Sleep Apnoea: a conceptual framework, In: Proc. CEUR Workshop(CI-Health 2012)Viola& Jones(2004) Robust Real-Time Face Detection. Int J Comput Vision ISO 14001"