Transcription of SeqFormer: a Frustratingly Simple Model for Video Instance ...
{{id}} {{{paragraph}}}
Example: confidence
SeqFormer: a Frustratingly Simple Model for Video Instance ...
SeqFormer: a Frustratingly Simple Model for Video Instance Segmentation Junfeng Wu1 Yi Jiang2 Wenqing Zhang1 Xiang Bai1 Song Bai2. 1 2. Huazhong University of Science and Technology ByteDance [ ] 15 Dec 2021. 60. Abstract SeqFormer 55. In this work, we present SeqFormer, a Frustratingly sim- YouTube-VIS 2019 AP. ple Model for Video Instance segmentation. SeqFormer fol- 50. Propose-Reduce lows the principle of vision transformer that models in- stance relationships among Video frames. Nevertheless, we 45 IFC. observe that a stand-alone Instance query suffices for cap- 40 VisTR. turing a time sequence of instances in a Video , but atten- CrossVIS. tion mechanisms should be done with each frame indepen- STEm-Seg 35. dently. To achieve this, SeqFormer locates an Instance in Mask-Track each frame and aggregates temporal information to learn 30.
the hand-crafted post-processing module. Deformable Detr [29] achieves better performance by using local attention and multi-scale feature maps. VisTR [24] is the first method that adapts DETR [4] to the VIS task and uses instance queries to obtain instance sequence from video clips. How-ever, VisTR can not handle variable-length or long-time
Loading..
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: