Shape description based on projected edges and global statistical features

Stubendek, Attila and Karacs, Kristóf and Roska, Tamás (2014) Shape description based on projected edges and global statistical features. In: International Symposium on Nonlinear Theory and its Applications (NOLTA 2014), September 14-18, 2014, Luzern, Svájc.

Text Shape_NOLTA14.pdf - Accepted Version Restricted to Registered users only Download (375kB)

Abstract

A projected edge based shape descriptor extended by global features is presented along with a related learning method. We also propose a two level classification method, corresponding the two distinct feature sets. Our experimental results show that the combination of independent features leads to increased recognition robustness and speed. The core algorithms are appropriate for cellular architectures and dedicated VLSI hardware.

Actions (login required)

Edit Item