search

carrierlxk / AGNN

Zero-shot Video Object Segmentation via Attentive Graph Neural Networks (ICCV2019 Oral)
207 stars 25 forks source link

N node graph vs N' node at implementation #18

Open ZahraPoorsoltani opened 1 year ago

ZahraPoorsoltani commented 1 year ago

hi, you say at paper that graph is N frame with fully connected, but why graph with 3 node should work at implemetation level?