SOTAVerified

Hyperbolic Image Embeddings

2019-04-03CVPR 2020Code Available1· sign in to hype

Valentin Khrulkov, Leyla Mirvakhabova, Evgeniya Ustinova, Ivan Oseledets, Victor Lempitsky

Code Available — Be the first to reproduce this paper.

Reproduce

Code

Abstract

Computer vision tasks such as image classification, image retrieval and few-shot learning are currently dominated by Euclidean and spherical embeddings, so that the final decisions about class belongings or the degree of similarity are made using linear hyperplanes, Euclidean distances, or spherical geodesic distances (cosine similarity). In this work, we demonstrate that in many practical scenarios hyperbolic embeddings provide a better alternative.

Tasks

Benchmark Results

DatasetModelMetricClaimedVerifiedStatus
CUB 200 5-way 1-shotHyperbolic ProtoNetAccuracy60.52Unverified
CUB 200 5-way 5-shotHyperbolic ProtoNetAccuracy72.22Unverified
Mini-Imagenet 5-way (1-shot)Hyperbolic ProtoNetAccuracy51.57Unverified
Mini-Imagenet 5-way (5-shot)Hyperbolic ProtoNetAccuracy66.27Unverified
OMNIGLOT - 1-Shot, 20-wayHyperbolic ProtoNetAccuracy95.9Unverified
OMNIGLOT - 1-Shot, 5-wayHyperbolic ProtoNetAccuracy99Unverified
OMNIGLOT - 5-Shot, 20-wayHyperbolic ProtoNetAccuracy98.15Unverified
OMNIGLOT - 5-Shot, 5-wayHyperbolic ProtoNetAccuracy99.4Unverified

Reproductions