Recently, deep learning applications to image compression and visual communication have attracted attention of many researchers. Applications to image compression can be classified to two methods: replacing a component of international standard video codec by deep learning (partial replacement method), and total codec implementation by deep learning (end-to-end method). Applications to visual communications can be classified to two research fields: estimation of subjective image quality by deep learning (picture quality assessment), and transmission rate control by predicting future throughputs by deep learning (adaptive rate control). This manuscript introduces recent trends of these researches, and also introduces research examples by the authors.

For the past few years, the trend of global warming has become evident than ever. The effect of global warming is expected to be enhanced in the cryosphere. Thus, importance of sea ice monitoring is increasing. Satellite observation is one of the most effective ways to monitor the distribution of sea ice on global basis. In this paper, the authors have examined the characteristics of various sensors onboard satellites for monitoring sea ice from space.