Salient regions detection using convolutional neural networks and color volume

2017 International Symposium on Application of Materials Science and Energy Materials

材料科学;能源学

Liu, Guang-Hai^1 ; Hou, Yingkun^2

College of Computer Science and Information Technology, Guangxi Normal University, Guilin, China^1

School of Information Science and Technology, Taishan University, Taian, SHANDONG

271000, China^2

关键词: Basic structure;    Computational burden;    Computing model;    Convolutional neural network;    Feature integration theories;    Saliency detection;    Salient regions;    State-of-the-art methods;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/322/7/072064/pdf
DOI  :  10.1088/1757-899X/322/7/072064

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

Convolutional neural network is an important technique in machine learning, pattern recognition and image processing. In order to reduce the computational burden and extend the classical LeNet-5 model to the field of saliency detection, we propose a simple and novel computing model based on LeNet-5 network. In the proposed model, hue, saturation and intensity are utilized to extract depth cues, and then we integrate depth cues and color volume to saliency detection following the basic structure of the feature integration theory. Experimental results show that the proposed computing model outperforms some existing state-of-the-art methods on MSRA1000 and ECSSD datasets.