Fig. 1. Proposed network framework

Fig. 2. Special Challenge Attribute Fusion Module

Fig. 3. Common Challenge Attribute Fusion Module

Fig. 4. Overall structure of the cross-modal interaction module

Fig. 5. Result of comparing our loss with binary loss

Fig. 6. Feature extraction results from the first stage

Fig. 7. Results of the evaluation on the GTOT dataset

Fig. 8. Evaluation results on the RGBT234 dataset

Fig. 9. Results of the evaluation on the Lasher dataset

Fig. 10. Visualization of tracking results for different frames under different video sequences in the Lasher datase