notesum.ai

Published at October 24

SegLLM: Multi-round Reasoning Segmentation

cs.CV

cs.AI

Released Date: October 24, 2024

Authors: XuDong Wang¹, Shaolun Zhang¹, Shufan Li², Konstantinos Kallidromitis³, Kehan Li⁴, Yusuke Kato³, Kazuki Kozuka³, Trevor Darrell¹

Aff.: ¹UC Berkeley; ²UCLA; ³Panasonic AI Research; ⁴UC Berkeley / Stanford

Arxiv: https://arxiv.org/abs/2410.18923v1

Refer to caption

Methods	RefCOCO				RefCOCO+				RefCOCOg
	val	testA	testB		val	testA	testB		val	test
VLT (Ding et al., 2021)	67.5	70.5	65.2		56.3	61.0	50.1		55.0	57.7
LAVT (Yang et al., 2022)	72.7	75.8	68.8		62.1	68.4	55.1		61.2	62.1
SEEM (Zou et al., 2023)	-	-	-		-	65.7	-		-	-
LISA-7B (Lai et al., 2024)	74.1	76.5	71.1		62.4	67.4	56.5		66.4	68.5
NExT-Chat (Zhang et al., 2024)	74.7	78.9	69.5		65.1	71.9	56.7		67.0	67.0
SegLLM (ours)	80.2	81.5	75.4		70.3	73.0	62.5		72.6	73.6