Siamese Network Based on Self-Attention Dropout and Blur Pooling for Object tracking

Key Points

• The aim is to enhance object tracking accuracy using a Siamese network by integrating self-attention dropout and blur pooling techniques.
• Developed a Siamese network architecture incorporating self-attention dropout.
• Implemented blur pooling to improve feature extraction.
• Evaluated performance on standard object tracking datasets.
• Achieved higher object tracking accuracy compared to traditional methods.
• Demonstrated robustness in challenging environmental conditions.