江苏师范大学贾志刚教授学术讲座通知
讲座时间：2025年12月16日（周二）上午10:30-11:30
讲座地点：广东技术师范大学工业中心802
报告题目：Interpretable feature modeling for robust color watermarking in the quaternion framework
主讲人简介：贾志刚
贾志刚，江苏师范大学数学与统计学院、数学研究院，教授、博导。2009年毕业于华东师范大学数学系，获理学博士学位；2023年入选江苏高校“青蓝工程”中青年学术带头人；2024年起担任学术期刊Numerical Algorithms的编委。主要研究方向为数值代数与图像处理，至今已在IEEE Trans. Image Process.，SIAM J. Matrix Anal. Appl., SIAM J. Sci. Comput., SIAM J. Imaging Sci.等期刊上发表学术论文50余篇，其中2 篇入选“ESI高被引”论文；在科学出版社（北京）出版英文专著1部（独立作者）；主持国家自然科学基金项目3项、省高校自然科学研究重大项目1项，参加国家自然科学基金重大项目和国家重点研发计划课题各1项；先后以第一完成人身份荣获第十届淮海科学技术奖（科技创新奖）一等奖和江苏省高等学校科学技术研究成果奖（自然科学奖）三等奖。曾到英国曼彻斯特大学、香港浸会大学、澳门大学等高校数学系进行学术访问。
报告摘要：In this talk, we systematically investigate the application of four quaternion matrix decompositions in blind color image watermarking. Two new forms of quaternion Givens transformation are proposed to overcome the difficulty of constructing numerical features in matrix decomposition for carrying watermarks, while further improving the efficiency and accuracy of computation. Next, four quaternion matrix decomposition algorithms are proposed based on quaternion Givens transformation, and their numerical features are specifically analyzed and matched with suitable watermarking mechanisms. As a result, four robust watermarking algorithms are proposed, while real structure-preserving arithmetic architectures are introduced to improve algorithm efficiency. Additionally, an image correction method is proposed to align the screenshot with the watermarked image geometrically, enhancing watermark extraction. Experimental outcomes reveal that, in comparison with other algorithms, the quaternion representation has a better reconstruction ability, the two new forms of the quaternion Givens transformation are also better in terms of efficiency and accuracy as well, and the four proposed color image watermarking algorithms show more satisfactory invisibility and robustness.
欢迎广大教师、科研人员、学生参加！
计算机科学学院 2025年12月12日
