HyperAI초신경

Image Matching On Imc Phototourism

평가 지표

mean average accuracy @ 10

평가 결과

이 벤치마크에서 각 모델의 성능 결과

		Paper Title
Harris Corner	0.65606	HarrisZ$^+$: Harris Corner Selection for Next-Gen Image Matching Pipelines
DISK	0.65435	DISK: Learning local features with policy gradient
SuperGlue	0.65248	SuperGlue: Learning Feature Matching with Graph Neural Networks
DoG-AffNet-HardNet8	0.64212	Repeatability Is Not Enough: Learning Affine Regions via Discriminability
Key.Net-SOSNet	0.60285	Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters
RootSIFT	0.59859	Three things everyone should know to improve object retrieval
R2D2	0.56345	R2D2: Reliable and Repeatable Detector and Descriptor
D2-Net (MS)	0.36285	D2-Net: A Trainable CNN for Joint Detection and Description of Local Features

0 of 8 row(s) selected.

Image Matching On Imc Phototourism | SOTA | HyperAI초신경