Abstract: In allusion to the some deficiencies from the algorithm based on error metrics
(QEM), such as neglect of some geometric features, a novel approach based on visual feature is
proposed for mesh simplification. The approach is driven by introducing a visual feature into the
new QEM to optimize error metrics, as well as half-edge collapse, therefore the collapse
sequences of edge can be adjusted, and some sharp visual features of the model can be preserved.
We define the visual feature of one vertex based on its vertex curvature entropy which reflects the
visual variation of the region centered at this vertex. Comparing with QEM algorithm, the
algorithm is efficient, reliable, and can maintain the visual characteristics of the model better.

Key words: mesh simplification, half-edge collapse, quadric error metrics, visual feature