self-attention MDNet

Self-Attention MDNet (SAM) is a deep learning framework for object tracking in videos. It is based on the popular MDNet architecture, which uses a siamese network to learn a similarity metric between the object of interest in the first frame and the subsequent frames. SAM extends MDNet by incorporating self-attention mechanisms, which allow the network to selectively attend to different parts of the input sequence based on their relevance to the task at hand. This helps the network to better handle occlusions, illumination changes, and other challenges that arise when tracking objects in videos. SAM has been shown to outperform state-of-the-art object trackers on several benchmark datasets.