notesum.ai

Published at November 9

Personalize to generalize: Towards a universal medical multi-modality generalization through personalization

cs.CV

cs.AI

Released Date: November 9, 2024

Authors: Zhaorui Tan¹, Xi Yang², Tan Pan¹, Tianyi Liu², Chen Jiang¹, Xin Guo¹, Qiufeng Wang², Anh Nguyen³, Yuan Qi¹, Kaizhu Huang⁴, Yuan Cheng¹

Aff.: ¹Shanghai Academy of Artificial Intelligence for Science; ²Xi'an Jiaotong-Liverpool University; ³University of Liverpool; ⁴Duke Kunshan University

Arxiv: http://arxiv.org/abs/2411.06106v1

	Target	T1			T1ce			T2			Flair
Source		PSNR↑	NMSE↓	SSIM↑	PSNR↑	NMSE↓	SSIM↑	PSNR↑	NMSE↓	SSIM↑	PSNR↑	NMSE↓	SSIM↑
	SwinUNETR	32.815	0.092	0.941	31.655	0.202	0.912	24.650	0.361	0.857	27.593	0.202	0.883
	MS-SPADE	29.001	0.055	0.942	26.119	0.078	0.912	25.818	0.103	0.904	24.842	0.113	0.859
T1	Ours	43.472	0.003	0.996	34.547	0.045	0.955	30.756	0.065	0.944	31.693	0.049	0.937
	SwinUNETR	32.456	0.100	0.929	33.001	0.156	0.926	25.125	0.366	0.859	27.699	0.211	0.882
	MS-SPADE	26.228	0.076	0.922	28.759	0.060	0.937	25.990	0.092	0.907	25.204	0.092	0.881
T1ce	Ours	34.077	0.020	0.962	46.663	0.003	0.996	30.775	0.063	0.942	32.224	0.046	0.941
	SwinUNETR	30.102	0.171	0.896	30.354	0.283	0.883	26.831	0.268	0.887	27.234	0.242	0.872
	MS-SPADE	25.422	0.085	0.908	25.234	0.087	0.895	29.230	0.048	0.942	25.074	0.098	0.867
T2	Ours	32.646	0.028	0.955	33.857	0.051	0.949	43.653	0.006	0.991	32.224	0.046	0.941
	SwinUNETR	31.371	0.135	0.916	31.285	0.240	0.905	25.579	0.338	0.867	29.092	0.148	0.923
	MS-SPADE	25.186	0.090	0.905	25.899	0.094	0.906	26.146	0.086	0.913	28.608	0.058	0.938
Flair	Ours	32.752	0.026	0.951	33.471	0.055	0.944	30.571	0.068	0.940	43.624	0.004	0.995