Artificial Intelligence Based Profession Prediction Using Facial Analysis
| dc.contributor.advisor | Anbarjafari, Gholamreza | |
| dc.contributor.advisor | Karabulut, Doğuş | |
| dc.contributor.author | Mert, Gülce Naz | |
| dc.date.accessioned | 2021-05-27T07:23:44Z | |
| dc.date.available | 2021-05-27T07:23:44Z | |
| dc.date.issued | 2020 | |
| dc.description.abstract | Youth unemployment is a global problem which affects millions of young people. One of the reasons for this is that young people are often misguided, or have adopted professions that are not a good fit for them. If an association between facial features and certain professions can be established using artificial intelligence, it is possible to guide young people into suitable career paths, providing them a better future with more satisfying jobs. In order to achieve this goal, different neural network models that employ deep learning and transfer learning were built, alongside with a dataset consisting of face images of people who are professionals in their fields. This data was then fed into these neural networks, testing effects of different networks and their parameters on the accuracy of predicting professions based on face images. The experiments however, did not lead to high accuracy rates. The results and networks are then analyzed and limitations are brought up. The possible solutions to what could have caused low accuracy rates are discussed. In estonian: Noorte tööpuudus on globaalne probleem mis mõjutab miljoneid noori. Üks põhjustest on kuna noori inimesi on tihti valesti juhitud või nad on omastanud ameteid mis pole neile sobilikud. Kui on võimalik leida assotsiatsioone näojoonte ja kindlate ametite vahel kasutades tehisintellekti, kas siis on võimalik juhtida noori inimesi parematele ametikohtadele, varustades neid parema tulevikuga, kus on rohkem rahuldavad töökohti. Et sellise saavutusega hakkama saada, ehitati erinevaid närvivõrgud mudeleid mis kasutavad süvaõpet ja ülekandmise õpe koos andmetega, mis koosnevad inimeste näo piltidest kes on oma ala professionaalid. See informatsioon siis sisestati närvi võrkudesse, katsetades erinevate võrkude efekte ja nende parameetreid näo järgi ameti valimise täpsuses. See katse kahjuks ei viinud kõrge täpsusega tulemusteni. Tulemused ja võrgud siis analüüsiti ja leiti limiidid. Võimalike lahendusi arutatakse selle üle mis võiksid tekitada vähese täpsusega tulemusi. | en |
| dc.identifier.uri | http://hdl.handle.net/10062/72059 | |
| dc.language.iso | eng | et |
| dc.rights | openAccess | et |
| dc.rights | Attribution-NonCommercial-NoDerivatives 4.0 International | * |
| dc.rights.uri | http://creativecommons.org/licenses/by-nc-nd/4.0/ | * |
| dc.subject | Machine Learning | en |
| dc.subject | Computer Vision | en |
| dc.subject | Deep Learning | en |
| dc.subject | Transfer Learning | en |
| dc.subject | masinõpe | et |
| dc.subject | arvuti nägemine | et |
| dc.subject | sügav õppimine | et |
| dc.subject | ülekandmise õpe | et |
| dc.title | Artificial Intelligence Based Profession Prediction Using Facial Analysis | en |
| dc.title.alternative | Tehisintellektipõhine elukutse ennustamine näoanalüüsi abil | et |
| dc.type | info:eu-repo/semantics/bachelorThesis | et |