Parallel border tracking in binary images for multicore computers

• DOI: 10.1007/s11227-023-05052-2
• Publication Date: 2023-01-23T09:08:40Z
• Authors: Victor M. Garcia-...; Pedro Alonso-Jordá
• Keywords: Multi-core processor; Tracking (education)

Abstract Border tracking in binary images is an important operation many computer vision applications. The problem consists finding borders a 2D image (where all of the pixels are either 0 or 1). There several algorithms available for this problem, but most them sequential. In former paper, parallel border algorithm was proposed. This designed to run Graphics Processing units, and it based on sequential known as Suzuki algorithm. we adapt previously proposed GPU so that can be executed multicore computers. resulting evaluated against its counterpart. results show performance worsens (or even fails) very large with borders. On other hand, efficiently cope images.