Texture analysis in watermarking paradigms

Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Конференция: 21st International Conference on Knowledge - Based and Intelligent Information and Engineering Systems, KES 2017

Год издания: 2017

Идентификатор DOI: 10.1016/j.procs.2017.08.019

Ключевые слова: arc, interpolation, keypoints, line segment, Local warping, motion level, video stabilization

Аннотация: Digital watermarking algorithms have been developed rapidly as a response on the challenges caused by various internet attacks that are distorted the content of the host image and watermark partially or fully. In this paper, the issues of texture analysis with a goal to detect the most suitable image areas for embedding are discussed. The statistical and model-based methods are investigated as a trade-off between the computational cost and quality of the detected areas, where the embedded bits of a watermark could be the most invisible for a human vision. The criteria for detection of such areas based on the textural, contrast, illumination, and color coherence of the host image and watermark are formulated. The experiments show that the statistical methods based on the gradient oriented Local Binary Patterns (LBP) provide better computational time regarding to fractal estimation of textural image areas. © 2017 The Author(s).

Ссылки на полный текст

Издание

Журнал: Procedia Computer Science

Выпуск журнала: Vol. 112

Номера страниц: 1460-1469

ISSN журнала: 18770509

Издатель: Elsevier B.V.

Авторы

  • Favorskaya M. (Siberian State Aerospace University, 31 Krasnoyarsky Rabochy ave., Krasnoyarsk, Russian Federation)
  • Pyataeva A. (Siberian Federal University, 26, Kirensky st., Krasnoyarsk, Russian Federation)
  • Popov A. (Siberian State Aerospace University, 31 Krasnoyarsky Rabochy ave., Krasnoyarsk, Russian Federation)

Вхождение в базы данных

Информация о публикациях загружается с сайта службы поддержки публикационной активности СФУ. Сообщите, если заметили неточности.

Вы можете отметить интересные фрагменты текста, которые будут доступны по уникальной ссылке в адресной строке браузера.