SALIENCY DETECTION VIA DENSE CONVOLUTION NETWORK
Journal Title: Topics in Intelligent Computing and Industry Design (ICID) - Year 2017, Vol 1, Issue 3
Abstract
Saliency detection is a fundamental problem in the field of image processing and computer vision. The convolutional model has been also used in saliency detection for its outstanding performance on image classification and localization task. In this paper, we propose a novel way to detect the salient object by modifying the Dense Convolution Network (DenseNet). We replace the fully-connected layer and the final pooling layer in DenseNet into a convolution layer and a deconvolution layer to fit the saliency detection task. And our network ends up with a squared Euclidean loss layer for saliency regression. Our network is end-to-end architecture which outputs saliency maps directly. Experimental results demonstrate that our approach is competitive in comparison with the state-of-the-art approaches.
Authors and Affiliations
Zheng Fang, Tieyong Cao, Yibo Xing
Keywords
Related Articles
- EP ID EP409155
- DOI 10.26480/iscsai.01.2017.33.35
- Views 80
- Downloads 0
THE INFLUENCE OF HORIZONTAL RING RIB ON VIBRATION AND NOISE OF BOX GIRDER IN URBAN RAILWAY TRANSIT
The box girder of urban rail transit is taken as the research object in this paper. By the combination of field measurement and numerical simulation, frequency spectrum of the vibration and radiated noise of the box gird...
ANALYSIS OF CONSUMPTION STRUCTURE FEATURED BY INBOUND TOURISTS IN ZHEJIANG PROVINCE
In accordance with the data related to the inbound tourism in Zhejiang province from 2006 to 2015, and based on the Excel software, the consumption structure of inbound tourists in Zhejiang province is analyzed. As, the...
APPLICATION OF FEEDFORWARD-FUZZY ADAPTIVE CONTROL IN THE CONTROL OF ADSORPTION TOWER OF DRY DESULFURIZATION
Flue gas purification of a domestic steel mill using activated carbon desulfurization process. Among them, the flue gas treatment method of the adsorption tower is filled with activated carbon and flue gas permeated and...
STUDY ON TOURIST SATISFACTION OF THE WAN-LV VALLEY RESORT BASED ON ONLINE COMMENTS
Using the six OTA platforms, we collect the visitors’ online comments on the Wan-lv valley resort and obtain valid data through identification, screening, induction and other means. Using ROST software analysis of high f...
COMPLICATED MECHANICAL EQUIPMENT DIAGNOSIS BASED ON BAYESIAN NETWORKS
Mechanical equipment fault diagnosis is a complicated process. Due to the complex structure, the different operating environment, the different detection means and testing equipment, the difference between the operator a...