Abstract: Existing light field salient object detection (LFSOD) models predominantly rely on convolutional neural networks or local attention to process light field data, consequently encountering ...