notesum.ai

Published at November 27

MvKeTR: Chest CT Report Generation with Multi-View Perception and Knowledge Enhancement

cs.CV

cs.AI

Released Date: November 27, 2024

Authors: Xiwei Deng¹, Xianchun He¹, Yudan Zhou¹, Shuhui Cai¹, Congbo Cai¹, Zhong Chen¹

Aff.: ¹Unknown

Arxiv: http://arxiv.org/abs/2411.18309v1

Refer to caption

All evaluation metrics are illustrated as percentage (%).
Method	Year	BLEU-1	BLEU-2	BLEU-3	BLEU-4	METEOR	ROUGE-L
Vanilla Transformer [21]^∗	2017	32.37	28.4	25.82	23.93	22.22	51.28
MRMA [31]^∗	2018	40.85	30.16	23.77	19.54	19.52	33.45
M2TR [32]^∗	2021	39.46	32.76	28.56	25.72	22.35	47.63
R2GenCMN [5]^∗	2022	48.29	38.34	32.53	28.42	23.89	48.91
M2KT [33]^∗	2023	39.71	33.17	29.12	26.16	20.11	47.19
TSGET [34]^∗	2024	46.12	38.15	32.92	28.84	23.02	50.19
UDT [35]^∗	2024	46.54	39.17	34.48	31.08	23.42	53.36
CAMANet [36]^∗	2024	54.28	45.26	39.42	35.38	26.48	54.2
CT2Rep [9]^∗	2024	49.81	40.7	35.19	31.28	24.82	51.18
SL-DG [8]^‡	2024	-	–	–	23.70	21.90	43.80
Dia-LLaMA [10]^‡	2024	51.16	–	–	29.64	26.28	42.15
MvKeTR(Ours)	–	58.36	48.79	42.43	37.86	28.36	54.25