HyperAI초신경

홈 뉴스 최신 연구 논문 튜토리얼 데이터셋 백과사전 SOTA LLM 모델 GPU 랭킹 컨퍼런스

한국어

HyperAI초신경

Image Super Resolution On Ffhq 256 X 256 4X

평가 지표

FID

MS-SSIM

PSNR

SSIM

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	FID	MS-SSIM	PSNR	SSIM	Paper Title	Repository
FSRCNN	139.78	0.930	22.45	0.709	Accelerating the Super-Resolution Convolutional Neural Network
CAGFace	74.43	0.958	27.42	0.816	Component Attention Guided Face Super-Resolution Network: CAGFace
BDPM	5.71	-	30.05	0.864	Binary Diffusion Probabilistic Model	-
SRFBN	132.59	0.895	21.96	0.693	Feedback Network for Image Super-Resolution
BRGM	-	-	24.16	0.70	Bayesian Image Reconstruction using Deep Generative Models
PULSE	-	-	15.74	0.37	PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models
EnhanceNet	116.38	0.897	23.64	0.701	EnhanceNet: Single Image Super-Resolution Through Automated Texture Synthesis
SRGAN	156.07	0.757	17.57	0.415	Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network
HiFaceGAN	5.36	0.971	28.65	0.816	HiFaceGAN: Face Renovation via Collaborative Suppression and Replenishment
SRCNN	147.21	0.900	23.12	0.688	Image Super-Resolution Using Deep Convolutional Networks
ESRGAN	166.36	0.747	15.43	0.267	ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks
EDSR	129.14	0.901	22.47	0.706	Enhanced Deep Residual Networks for Single Image Super-Resolution

0 of 12 row(s) selected.