基於分析手機感測器資料美化手寫軌跡

Abstract

在現今的資訊時代中，人們使用智慧型手機的情況已經是相當普遍。而智慧型手機內部有許多種類的感測器，這些感測器可能蘊藏了與人們使用智慧型手機的行為相關的資訊。而近年來，人們於智慧型手機觸控螢幕輸入手寫軌跡訊息是相當普遍的，而且人們也時常在不同情境下使用智慧型手機(舉個例子來說，走路、慢跑、坐公車. . .等等)。當人們在不同情境下使用智慧型手機輸入手寫軌跡訊息，可能會受到當下情境的影響造成其手寫軌跡訊息不美觀，我們希望利用我們現有的視覺化工具分析感測器與人們不美觀的手寫軌跡訊息之間的關係趨勢。我們也將我們現有的視覺化工具的呈現方式實作在智慧型手機上，幫助我們在不同情境下即時的觀察手寫軌跡資訊與手機感測器之間的關係。我們也實際在晃動情境下蒐集使用者手寫軌跡資料(舉個例子來說，坐公車的時候)，藉由當下情境與統整後的觀察，分析手機感測器與觸控位置偏差之間的關係趨勢，進而設計平滑方法和回歸模型，達到手寫美化的效果。

In modern information age, smartphones are common commodities for most people. There are many sensors in the smartphone, these sensors store useful information between people’s behavior and smartphones. And in recent years, smartphone handwriting trajectory input methods are more common. Smartphone users are using their phones in different situations(For examples, walking, jogging, riding the bus, etc.). When people use smartphones to enter handwriting trajectory information, the unstable situations may cause difficulty in processing these handwriting trajectory information. We hope to use our existing visualization tool to analyze the relation between sensors and unrecognizable handwriting trajectory information. To help us observe the relation between handwriting trajectory information and smartphone sensor data immediately, we also implement the presentation of our existing visualization tool on smartphones. In addition, we collect the handwritten user data of users in the unsteady situations. Base on the observations of the environment and processed data, we can present the trend of the relation between the smartphone sensor and the touch position. We demonstrate our smooth method and regression model to beautify handwriting.