Frequency Salient Based Fiber Detection
Yang Liu, Duo Zhang, Chunlei Li, Zhenduo Guo, Chaodie Liu · 2016
The use of random and unique characteristics of the fiber distribution is an effective fiber anti-fake method in modern industry.Detecting and combing with the distribution of the fiber from images automatically can improve the efficiency of industrial production greatly.In this paper, we propose a frequency salient based fiber detection method which is robust to light condition and slightly background impurity.Our method exploits low level features of color and luminance, which is easy to implement, fast, and provides full resolution.The resulting saliency maps are better suited to fiber detection.The experiments are conducted on our released database, which include more than 10,000 fiber images, and the experimental results demonstrate the effectiveness of the proposed method.