prompt

[lmr2706]

Hello, I would like to know specifically how you leverage multiple prompts to guide the segmentation, and how these different prompts coordinate with each other to achieve the best performance. Could you also explain how the prompts are obtained in both the training and testing stages? Thank you.

[lmr2706]

Could you tell me how you incorporate prompts into the interaction map, how the interaction map is obtained, in what form the mask prompt for the third channel exists, and how the prompts across these three channels interact with each other to function?