Skip to main content

Hitachi

Corporate InformationResearch & Development

June 10, 2014

Report from Presenter


Photo 1 Presentation scene



Fig.1 System image
Enlarge



Photo 2 Experiment scene
Note that the photo used in the presentation was
replaced.

The International Conference on Information Science, Electronics and Electrical Engineering (ISEEE 2014) was held in the city of Sapporo, Japan from April 26 - 28, 2014 at Sapporo Prince Hotel. About 180 participants, who were practitioners and academic researchers, came from countries such as China, Europe, USA, and Japan. They discussed heavily various topics such as energy-efficient system techniques, information security, and chip design.

Hitachi's Yokohama Research Laboratory made a presentation titled "Use of Rough Image Analysis in Smartphone Application for Video Watermark Detection," co-authored with Hitachi Solutions, Ltd. and Japan Broadcasting Corporation.

A method has been developed for image analysis used for efficiently detecting watermarks in video content displayed by using a smart phone. A person viewing a video with watermarks could use a smartphone equipped with our reader application to capture frame images. The application would then be used to detect the watermarks in each frame image and then provide information corresponding to the video (see Fig. 1).

The computational load involved in processing a full-size camera image is reduced by using Gaussian filtering to roughly analyze the complexity of the image regions. Since watermarks are assumed to be robustly embedded in those regions with higher complexity, each region is checked for watermarks in descending order of complexity. Quick and reliable watermark detection is achieved in real time by adjusting the number of checked regions so that the calculation speed matches device-hardware performance.

System testing showed that our reader application with the developed method is practical to some degree under the constraint that the camera is positioned within a particular range with respect to the displayed video. Just like a barcode, almost-imperceptive watermarks can be used for various services by linking cyber and physical space.

At the conference, I had valuable discussions with practitioners and gained valuable knowledge such as on up-to-date implementation techniques for smart devices and efficient coding methods for images.

(By YAMADA Takaaki)

Related Papers

  • Takaaki Yamada, Motoki Kamitani, Haruyuki Nakamura, and Tsuyoshi Ohta: "Use of Rough Image Analysis in Smartphone Application for Video Watermark Detection", in Proc. of Int'l Conf. on Information Science, Electronics and Electrical Engineering (ISEEE 2014), pp. 474-478, 2014.
  • Page top