Image cross-modal saliency detection method and device, equipment and medium
Patent Information
- Application Number
- CN202511584168.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-31
- Publication Date
- 2026-01-30
AI Technical Summary
Existing neural network-based salient object detection methods have low accuracy and lack interpretability, making it difficult to meet the high-level semantic and abstract knowledge requirements in multimodal image processing.
By acquiring color and depth images, a first phase map and a second phase map are determined respectively. The saliency phase map is iteratively updated using a phase update and transformation model. Combined with the conversion between the frequency domain and the spatial domain, a saliency map is generated.
It achieves cross-modal saliency detection with high accuracy and interpretability, and can effectively utilize information from both color and depth modalities to generate saliency maps that conform to human visual attention mechanisms.
Smart Images

Figure CN121437852A_ABST
