稀疏Transformer | ViLab https://vilab.team/tag/%E7%A8%80%E7%96%8Ftransformer/ 稀疏Transformer Hugo Blox Builder (https://hugoblox.com)en-usSun, 16 Jun 2024 00:00:00 +0000 https://vilab.team/media/icon_hu2896232876136423579.png 稀疏Transformer https://vilab.team/tag/%E7%A8%80%E7%96%8Ftransformer/ Scene adaptive sparse transformer for event-based object detection https://vilab.team/publication/scene-adaptive-sparse-transformer-for-event-based-object-det/ Sun, 16 Jun 2024 00:00:00 +0000 https://vilab.team/publication/scene-adaptive-sparse-transformer-for-event-based-object-det/ <p>本文针对事件相机目标检测中Transformer计算成本过高的问题，提出场景自适应稀疏Transformer（SAST）。该方法通过窗口-令牌协同稀疏化与场景特定稀疏优化，在保持低计算量的同时实现高检测性能，并能根据场景复杂度自适应调整稀疏程度，有效平衡了检测精度与效率。</p>