Salient Object Detection Based on Optimization of Feature Computation by Neutrosophic Set Theory

Area of Science:

• Computer Vision
• Artificial Intelligence
• Image Processing

Background:

• Current saliency detection methods often struggle with feature selection and saliency map detail processing.
• This leads to degraded performance in detecting salient objects.

Purpose of the Study:

• To propose an improved salient object detection method using neutrosophic set (NS) theory.
• To address limitations in feature utilization and saliency map detail refinement.

Main Methods:

• Building prior object knowledge using foreground and background models (pixel-wise and super-pixel cues).
• Selecting and extracting feature maps for computation to separate object and background features.
• Fusing low-rank matrix recovery model features with object prior knowledge.
• Developing a novel mathematical description of neutrosophic set theory for saliency detection.

Main Results:

• The proposed method demonstrates competitive and superior results compared to state-of-the-art methods.
• Experiments on five public datasets validate the effectiveness of the approach.

Conclusions:

• The neutrosophic set theory-based salient object detection method effectively optimizes features and refines saliency map details.
• This approach offers improved accuracy and performance in salient object detection tasks.