Image based solar position editing and relighting

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2018

Öğrenci: MURAT TÜRE

Danışman: AHMET OĞUZ AKYÜZ

Özet:

Capturing perfect photos is a challenging task that requires good composition andtiming. Failure to achieve them may result in ‘missing the moment’. For example,when shooting photos at sunset the sun quickly disappears behind the horizon. Inthis thesis, a new image editing algorithm is defined to let the user change the sunposition in a photo. This is the first study in the literature which tries to solve this par-ticular image editing problem. The algorithm can be divided into three parts. First,the image pixels are segmented into different categories, namely the sky, clouds, sun,horizon, and the foreground. Then, a real-time rendered sky is replaced with the oldsky to allow changing the position of the sun. A color transfer algorithm is used to cope with the differences that cannot be accurately deduced from a single photo suchas exposure, atmospheric parameters, and the camera response curve. Furthermore,techniques for relighting the cloud and foreground pixels are proposed. Finally, theserelighted pixels are blended together by using the labels from the segmentation partof the algorithm to generate the final image. The proposed image editing technique works in real-time to provide instant feedback to the user. It also provides intutiveuser controls to allow overriding and fine-tuning various automatically detected parameters.