An Improved Ft Algorithm Based on Center Enhancement and Eigenvalue Normalization
Qi Zhang, Baolong Guo, Zuming Chen, Cheng Li · Journal of Physics Conference Series · 2018
The Frequency-tuned method (FT) detection does not work very well when the saliency maps is large and the background environment is complex. Aiming to solve the above problems, this paper improves the traditional FT method, which introduces centre enhancement and eigenvalue normalization. The improvements are implemented to three aspects: enhancing the central area, normalizing the LAB colour feature values, and weighting the LAB three-channel information. Experimental results show that the algorithm is superior to the origin FT in the performance of significant detection, accuracy, recall and other performance.