Saliency Assessment Using Selective Features Based on Entropy and Wavelet Transform
Bachir Kaddar, Hadria Fizazi, Dou El Kefel Mansouri · 2019
In this paper, we present a novel framework for an image saliency assessment based on an adaptive model. This later evaluates the image-content importance using a tuning strategy based on information theoretic concepts coupled with wavelet multiscale image representation. Our saliency-based encoding form can greatly characterize both regular/irregular structures within the image. The performance of the proposed model is benchmarked with those available in the literature. The saliency assessment mechanism we propose can potentially offer up to very promising results.