AI Paper Atlas

Research map / Continual learning / Class-incremental

iNeMo: Incremental Neural Mesh Models for Robust Class-Incremental Learning

Tom Fischer, Yaoyao Liu, Artur Jesslen, Noor Ahmed, Prakhar Kaushik, Angtian Wang, Alan Yuille, Adam Kortylewski, Eddy Ilg

ECCV 2024 · Continual learning › Class-incremental · approach: Exemplar-efficient class-incremental learning · 7 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 pose-aware exemplar selection instead of herding (ablation)iNeMo · P3D B0+3final accuracy: 93.60 -> 94.70Improved
add latent space initialization (ablation)iNeMo · O3D B0+20final accuracy: 70.32 -> 71.67Improved
add positional regularization (ablation)iNeMo · P3D B0+3final accuracy: 94.78 -> 94.98Improved
add knowledge distillation to the full latent-space setup (ablation)iNeMo · O3D B0+20final accuracy: 72.09 -> 80.17Improved
add uncertainty term to classification inference (ablation)iNeMo · O3D B0+20final accuracy: 75.8 -> 80.17Improved
fine-tune NeMo incrementally without replay or distillation (ablation)NeMo · P3D B0+3final accuracy: 17.47Hurt
apply LwF to NeMo instead of ResNet50 (ablation)LwF · P3D B0+3final accuracy: 83.75 -> 17.45Hurt
add exemplar replay (iCaRL herding) to NeMo (ablation)NeMo · P3D B0+3final accuracy: 17.45 -> 93.72Improved
reduce exemplars per class from 20 to 5 (ablation)iNeMo · P3D B0+3final accuracy: 96.41 -> 82.59Hurt
add pose supervision head to ResNet50 baseline (ablation)iCaRL ResNet50 · P3D B0+3final accuracy: 91.79 -> 89.86Hurt
use NeMo representation instead of ResNet50 for incremental classificationiNeMo vs 2D baselines · Occluded-P3D L3accuracy: 55.89 -> 71.55Improved
iNeMo vs 2D baselines for pose pi/6 under strong corruptioniNeMo · Corrupted-P3D pose pi/6accuracy: 47.30 vs 54.57Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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