ScanShot: Detecting Document Capture Moments and Correcting Device Orientation

ScanShot: Detecting Document Capture Moments and Correcting Device Orientation
Jeungmin Oh, Woohyeok Choi, Joohyun Kim, Uichin Lee

CHI '15: ACM Conference on Human Factors in Computing Systems
Session: Phones for more than Just Talking & Text

Abstract
Document capturing with smartphone cameras is performed increasingly often in our daily lives. However, our user study results (n=10) showed that more than 80% of landscape tasks had incorrect orientations. To solve this problem, we systematically analyzed user behavior of document capturing and proposed a novel solution called ScanShot that detects document capturing moments to help users correct the orientation errors. ScanShot tracks the gravity direction to capture document capturing moments, analyzes logged gyroscope data to automatically update orientation changes, and provides visual feedback of the inferred orientation for manual correction. Our user study results (n=20) confirmed that capturing moments can be recognized with accuracy of 97.5%, our update mechanism can reduce the orientation errors by 59 percentage points.

Recorded at the 33rd Annual ACM Conference on Human Factors in Computing Systems in Seoul, Korea, April 18-23, 2015