报告人：罗守胜（浙江师范大学）

邀请人：刘海霞

报告时间：2023年12月20日（星期三）10：00-12:00

报告地点：腾讯会议：288 410 585

报告题目：A new boundary-based method for convexity characterization using binary function    

报告摘要：In real images, the integrities of objects are destroyed by occlusion, illumination bias and noise. Convexity of shapes is one of very important cues for human vision system. Therefore, it is important to incorporate the convexity of shapes for image processing (e.g. segmentation). The characterization of convexity plays a very important role in real implementation. Existing characterization methods can be grouped in boundary-based methods and region-based methods. Roughly speaking, level set function (resp. binary function) is used to represent concerned objects for the boundary-based (resp. region-based) methods. In this talk, I will present a new binary-based method for convex object characterization using binary function rather than level set function after reviewing the existing characterizations. Applications for image segmentation are demonstrated.

报告人简介：罗守胜，浙江师范大学副教授。 2013 年博士毕业于北京大学数学学院， 2017-2019 年在北京计算科学研究中心/香港科技大学做联合博士后研究，曾访问佐治亚理工大学数学系、香港浸会大学数学系。主要研究方向为 CT 图像重建、带有凸先验的图像分割，在 IPI、 JMIV、 ICCV 等期刊和会议上发表论文多篇，主持和参与国家自然科学基金多项。