SOTAVerified

Novel Class Discovery

The goal of Novel Class Discovery (NCD) is to identify new classes in unlabeled data, by exploiting prior knowledge from known classes. In this specific setup, the data is split in two sets. The first is a labeled set containing known classes and the second is an unlabeled set containing unknown classes that must be discovered.

Papers

Showing 1120 of 65 papers

TitleStatusHype
Bootstrap Your Own Prior: Towards Distribution-Agnostic Novel Class DiscoveryCode1
Bridging the Gap: Learning Pace Synchronization for Open-World Semi-Supervised LearningCode1
Class-incremental Novel Class DiscoveryCode1
Class-relation Knowledge Distillation for Novel Class DiscoveryCode1
Dual-level Adaptive Self-Labeling for Novel Class Discovery in Point Cloud SegmentationCode1
A Unified Objective for Novel Class DiscoveryCode1
AutoNovel: Automatically Discovering and Learning Novel Visual CategoriesCode1
Encoding Structure-Texture Relation with P-Net for Anomaly Detection in Retinal ImagesCode1
Enhancing Novel Object Detection via Cooperative Foundational ModelsCode1
Meta Discovery: Learning to Discover Novel Classes given Very Limited DataCode1
Show:102550
← PrevPage 2 of 7Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1AutoNovelClustering Accuracy0.92Unverified
#ModelMetricClaimedVerifiedStatus
1AutoNovelClustering Accuracy0.75Unverified
#ModelMetricClaimedVerifiedStatus
1AutoNovelClustering Accuracy0.95Unverified