Main subject detection via adaptive feature refinement

Cuong Vu

Journal of Electronic Imaging · 2011 · 20 citations · 32 references

Concepts

Abstract

In this paper we present an algorithm which uses adaptive selection of low-level features for main subject detection. The algorithm first computes low-level features such as con-trast and sharpness, each computed in a block-based fashion. Next, the algorithm quantifies the usefulness of each feature by using both statistical and geometric information measured across blocks. Finally, the saliency of each block is deter-mined via a weighted linear combination of the features, where the weights are chosen based on each feature’s esti-mated usefulness. Our results demonstrate that the adaptive nature of this algorithm allows it to perform competitively with other techniques, while maintaining very low computa-tional complexity. Index Terms — Main subject detection, low-level feature, adaptive feature selection, block-based. 1.

References

32