[2K4-IS-1a] Knowledge engineering

[2K4-IS-1a-05] Location Awareness Google Street Image Retouching with Seasonal Changes

〇Chiau Wei Huang¹, Ying Yu Chen¹, Hao Ming Hung¹, Hsiao Kuang Wu², Lieu Hen Chen¹ (1. National Chi Nan University, Taiwan, 2. National Central University, Taiwan)

Keywords:Deep Learning, Live Street Webcam Images, Street Images, Seasonal Changes, Smart City

Streetscape, which represents the characteristics of a city, serves as a canvas that reflects the cultural essence and the emotional connection of its residents. Among its elements, street trees, in particular, illustrate the cyclical rhythm of nature and time through their seasonal changing appearances. In this paper, we propose a location aware system which enables users to experience the seasonal growth stages of street trees within the images of Google Street View. Our system integrates deep learning, computer graphics, and web data mining technologies. Tree images captured from Google MAP are dynamically retouched based on their GPS information and our database, which has web-crawled daily YouTube live camera data since 2021. The experimental results shown that our method can generate vivid, engaging, and dynamic images of street trees that align with the local natural and environmental conditions, therefore successfully enhancing users' experience and perception of seasonal changes in the street images.