AI Paper Atlas

Research map / Continual learning / Online and replay-based

Continual Learning for Image-Based Camera Localization

Shuzhe Wang, Zakaria Laskar, Iaroslav Melekhov, Xiaotian Li, Juho Kannala

ICCV 2021 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks · 32 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add coverage-score buffering instead of class-balance bufferingHSCNet continual learning · i7S buffer 256 Img-buffpose accuracy %: 70.06 -> 72Improved
add coverage-score buffering instead of reservoir samplingHSCNet continual learning · i12S buffer 256 Img-buffpose accuracy %: 82.54 -> 91.85Improved
add coverage-score buffering instead of class-balance bufferingHSCNet continual learning · i19S buffer 256 Img-buffpose accuracy %: 64.7 -> 68Improved
add coverage-score buffering instead of class-balance bufferingHSCNet continual learning · i19S buffer 128 Img-buffpose accuracy %: 47.98 -> 46.43Hurt
add coverage-score buffering instead of class-balance buffering, 5-seed runsHSCNet continual learning · i7S buffer 256 Img-bufffinal accuracy %: 69.66 ± 0.54 -> 70.51 ± 0.94Improved
enlarge buffer from 128 to 1024 for Buff-CS coverage score on i19S (ablation)Buff-CS · i19Scoverage score %: 58.8 -> 96Improved
enlarge buffer from 128 to 256 for Buff-CS (ablation)Buff-CS · i19S Img-buffpose accuracy %: 46.43 -> 68Improved
store intermediate representations (Rep-buff) in addition to images (ablation)Buff-CS · i19S buffer 128pose accuracy %: 46.43 -> 51.86Improved
store intermediate representations (Rep-buff) instead of images only (ablation)Class-balance · i19S buffer 128pose accuracy %: 47.98 -> 54.12Improved
store intermediate representations (Rep-buff) instead of images only (ablation)Buff-CS · i7S buffer 256pose accuracy %: 72 -> 71.34No clear effect
add any replay buffer versus training without buffering (ablation)HSCNet continual learning · i7S buffer 256pose accuracy %: 10.61 -> 72.00Improved
Rep-buff storage compared to Img-buff (ablation)Buff-CS · buffer 256 memorystorage (Mb): 117 -> 1091Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy