HyperAI초신경

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

한국어

HyperAI초신경

Deblurring On Based

평가 지표

ERQAv2.0

LPIPS

PSNR

SSIM

Subjective

VMAF

평가 결과

이 벤치마크에서 각 모델의 성능 결과

							Paper Title	Repository
NAFNet (REDS)	0.74508	0.08561	30.54803	0.95035	2.8405	66.85941	Simple Baselines for Image Restoration
VRT (GoPro)	0.74874	0.08165	31.42945	0.94503	2.3854	66.72253	VRT: A Video Restoration Transformer
VRT (REDS)	0.75056	0.08248	30.97878	0.94601	1.5660	66.81782	VRT: A Video Restoration Transformer
DeblurGAN Inception	-	-	-	-	1.0375	-	DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better
MAXIM (REDS)	0.74277	0.07836	-	0.94959	1.0081	67.3502	MAXIM: Multi-Axis MLP for Image Processing
Deeprft (GoPro)	0.74323	0.08326	31.57612	0.94484	0.5354	66.55057	Intriguing Findings of Frequency Selection for Image Deblurring
Deeprft (REDS)	0.74339	0.08139	31.32349	0.94479	0.4622	66.46811	Intriguing Findings of Frequency Selection for Image Deblurring
MPR local	0.74521	0.08323	31.65037	0.94542	0.4407	67.01788	Improving Image Restoration by Revisiting Global Information Aggregation
MAXIM (GoPro)	-	0.09188	31.36344	0.94386	0.2070	67.7557	MAXIM: Multi-Axis MLP for Image Processing
Restormer local	0.73875	0.08251	31.12341	0.94217	0.1231	65.25911	Restormer: Efficient Transformer for High-Resolution Image Restoration
Restormer	0.74776	0.08239	31.76111	0.94632	0.1175	66.3964	Restormer: Efficient Transformer for High-Resolution Image Restoration

0 of 11 row(s) selected.