行動OLED顯示器針對應用於即時攝影的節能技術

Abstract

隨著OLED行動裝置迅速發展，如今行動裝置搭配OLED顯示器的情況已越來越受到市場青睞。但是在行動裝置電量有限的情況下，OLED顯示器仍然是最耗電的元件之一。因此也越來越多人開始針對OLED顯示器進行了電源管理，造就了許多OLED顯示器在省電議題上的相關研究。本論文探討了稱之為CURA框架的圖像像素縮放技術，基於該框架我們提出了Fast-CURA加速策略，該策略引入了一個新的視覺注意力模型，還用到的平行運算和機器學習的技術，目的是讓影像分析的執行時間縮短，讓Fast-CURA能夠推廣到需要即時處理的應用上。因此我們在搭載Android系統的智慧型手機上，開發了一個透過手機攝像機即時攝影的應用程式，該程式能夠錄製在OLED顯示器中播放的節能影片，降低影片播放時的功率消耗。經實驗證明該策略保留了CURA在視覺體驗及手機省電的效果，而且還讓執行時間變快了37至47.1倍。With the rapid development of OLED mobile devices, the situation of mobile devices with OLED displays has become more and more popular with the market. However, in the case of limited power consumption of mobile devices, OLED displays are still one of the most power-consuming components. Therefore, more and more people have begun to carry out power management for OLED displays, resulting in many OLED displays related research on power saving issues. This paper discusses the image pixel scaling technology called CURA framework. Based on this framework, we propose the Fast-CURA acceleration strategy. This strategy introduces a new visual attention model, and also uses parallel computing and machine learning. Technology, the purpose is to shorten the execution time of image analysis, so that Fast-CURA can be extended to applications that require immediate processing. Therefore, we have developed an application for real-time photography through a mobile phone camera on a smartphone equipped with an Android system. This application can record energy-saving videos played on an OLED display and reduce power consumption during video playback. Experiments prove that the strategy retains the effect of CURA in visual experience and mobile phone power saving, and also increases the execution time by 37 to 47.1 times.