Textonboost based on differential evolution
Keiko Ono, Daisuke TAWARA, Yoshiko Hanada · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2019
Semantic segmentation on a pixel basis is necessary for the semantic understanding of an image. Although the use of CNN is mainstream in the case where there are sufficient test images, in this research we aim to develop a method that is robust in an environment with few test images. Specifically, we improve Textonboost by using a differential evolution method.