Home / Current Issue / Paper 1715817
Confidence-Guided Dual Stream Hybrid Network for Salient Object Detection
Subject area: Science,Engineering and Technology · Area of research: Computer Vision and Deep Learning
DOI: https://doi.org/10.64388/IREV9I9-1715817
Abstract
Salient Object Detection (SOD) aims to automatically identify visually important regions in an image and is considered a fundamental task in computer vision. It plays an important role in various applications such as image segmentation, object recognition, visual tracking, and scene understanding. However, variations in object appearance, complex backgrounds, low contrast between foreground and background, and the presence of multiple objects make automatic saliency detection a challenging problem. Traditional saliency detection methods often fail to effectively capture both global contextual information and fine spatial details, which leads to incomplete object detection and inaccurate saliency maps. To overcome these limitations, this paper proposes a Confidence-Guided Dual-Stream Network for efficient salient object detection. The proposed framework utilizes two parallel feature extraction streams to capture complementary visual representations from input images. In addition, a confidence-guided mechanism evaluates the reliability of extracted features and adaptively adjusts their contributions during feature fusion. Experimental results on benchmark datasets demonstrate that the proposed approach produces more accurate saliency maps and improves salient object localization in complex visual scenes.
Keywords
Computer Vision, Confidence-Guided Feature Fusion, Deep Learning, Dual-Stream Network, Image Processing, Salient Object Detection.
How to cite this paper
@article{1715817,
author = {Dr. B. Hemantha Kumar, Rohitha Maddineni, Karanam Ramyasri, Gogineni Charan Sai, Borra Kasi Naga Sai Lakshmi},
title = {Confidence-Guided Dual Stream Hybrid Network for Salient Object Detection},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {9},
pages = {3204-3211},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1715817.pdf},
abstract = {Salient Object Detection (SOD) aims to automatically identify visually important regions in an image and is considered a fundamental task in computer vision. It plays an important role in various applications such as image segmentation, object recognition, visual tracking, and scene understanding. However, variations in object appearance, complex backgrounds, low contrast between foreground and background, and the presence of multiple objects make automatic saliency detection a challenging problem. Traditional saliency detection methods often fail to effectively capture both global contextual information and fine spatial details, which leads to incomplete object detection and inaccurate saliency maps. To overcome these limitations, this paper proposes a Confidence-Guided Dual-Stream Network for efficient salient object detection. The proposed framework utilizes two parallel feature extraction streams to capture complementary visual representations from input images. In addition, a confidence-guided mechanism evaluates the reliability of extracted features and adaptively adjusts their contributions during feature fusion. Experimental results on benchmark datasets demonstrate that the proposed approach produces more accurate saliency maps and improves salient object localization in complex visual scenes.},
keywords = {Computer Vision, Confidence-Guided Feature Fusion, Deep Learning, Dual-Stream Network, Image Processing, Salient Object Detection.},
month = {March},
doi = {https://doi.org/10.64388/IREV9I9-1715817}
}