A novel fully convolutional network for visual saliency prediction

Bashir Muftah Ghariba; Mohamed S Shehata; Peter McGuire

doi:10.7717/peerj-cs.280

A novel fully convolutional network for visual saliency prediction

PeerJ Comput Sci. 2020 Jul 13:6:e280. doi: 10.7717/peerj-cs.280. eCollection 2020.

Authors

Bashir Muftah Ghariba^{1

2}, Mohamed S Shehata³, Peter McGuire⁴

Affiliations

¹ Faculty of Engineering & Applied Science, Memorial University of Newfoundland, St. John's, NL, Canada.
² Department of Electrical and Computer Engineering, Faculty of Engineering, Elmergib University, Khoms, Libya.
³ Department of Computer Science, Mathematics, Physics and Statistics, University of British Columbia, Kelowna, BC, Canada.
⁴ C-CORE, St John's, NL, Canada.

Abstract

A human Visual System (HVS) has the ability to pay visual attention, which is one of the many functions of the HVS. Despite the many advancements being made in visual saliency prediction, there continues to be room for improvement. Deep learning has recently been used to deal with this task. This study proposes a novel deep learning model based on a Fully Convolutional Network (FCN) architecture. The proposed model is trained in an end-to-end style and designed to predict visual saliency. The entire proposed model is fully training style from scratch to extract distinguishing features. The proposed model is evaluated using several benchmark datasets, such as MIT300, MIT1003, TORONTO, and DUT-OMRON. The quantitative and qualitative experiment analyses demonstrate that the proposed model achieves superior performance for predicting visual saliency.

Keywords: Convolutional neural networks; Deep learning; Encoder-decoder architecture; Fully Convolutional Network; Human eye fixation; Semantic Segmentation.

Grants and funding

Bashir Muftah Ghariba received financial support from the Libyan Ministry of Higher Education and Scientific Research, and Elmergib University, Alkhums, for the PhD program. Memorial University of Newfoundland supported the publication fee. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.