最新ECCV 2026录用论文，探索如何对预训练视觉Transformer的表示进行可控操作，为视觉AI带来新的灵活性。
arXiv:2604.02327v2 Announce Type: replace Abstract: Pretrained Vision Transformers (ViTs) such as DINOv2 and MAE provide generic image features that c…
融合拓扑数据分析与视觉Transformer，在脑肿瘤四分类任务上取得新突破。
arXiv:2606.00927v1 Announce Type: new Abstract: Accurate brain tumor classification from magnetic resonance imaging (MRI) is a key requirement for ear…
提出测试时训练新范式，将Vision Transformer复杂度从二次降至线性，推理时仅需少量计算即可大幅提升准确率和鲁棒性。
arXiv:2605.02772v2 Announce Type: replace Abstract: While linear-complexity attention mechanisms offer a promising alternative to Softmax attention fo…
通过修改训练协议而非架构，让Vision Transformer更好地利用空间局部性，提升性能与效率。
arXiv:2605.16390v1 Announce Type: cross Abstract: We investigate whether the training protocol can induce spatial locality in the early layers of a Vi…