AI Paper Atlas

Research map / Detection and segmentation / 3D detection and driving

Disentangling 3D Prototypical Networks for Few-Shot Concept Learning

Mihir Prabhudesai, Shamit Lal, Darshan Patil, Hsiao-Yu Tung, Adam W Harley, Katerina Fragkiadaki

ICLR 2021 · Detection and segmentation › 3D detection and driving · approach: Unified LiDAR 3D detection · 2 citations (OpenAlex)

Read the paper

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
use disentangled 3D shape and style prototypes instead of the 3DP-Net representation without object disentanglementfew-shot concept classification · CLEVR, Real Veggie, and ReplicaaccuracyImproved
use D3DP-VQA rather than a 2D NSCL modelCLEVR VQA · one-shot novel-object test, 10 training examplesaccuracy: 0.594 -> 0.775Improved
use 2D NSCL rather than D3DP-VQACLEVR VQA · in-domain test, 250 training examplesaccuracy: 0.990 -> 0.939Hurt
add 3D shape prototypes instead of spatially pooled 1D shape codes (ablation)D3DP-VQA · CLEVR VQA, one-shot novel-object test, 10 training examplesaccuracy: 0.410 -> 0.775Improved
add shape/style disentanglement (ablation)D3DP-VQA · CLEVR VQA, one-shot novel-object test, 10 training examplesaccuracy: 0.457 -> 0.775Improved
add both 3D shape prototypes and shape/style disentanglement (ablation)D3DP-VQA · CLEVR VQA, one-shot novel-object test, 10 training examplesaccuracy: 0.608 -> 0.775Improved
replace disentangled features with entangled features (ablation)CLEVR VQA · one-shot novel-object test, 10 training examplesaccuracy: 0.619 -> 0.775Hurt
use InstanceNorm disentangled features with a rotation-aware check rather than the full method (ablation)CLEVR VQA · one-shot novel-object test, 250 training examplesaccuracy: 0.845 -> 0.861Improved
add generated imagined scenes to real detector training data3D object detector · CARLA, 50 annotations, IoU .3mean average precision: 0.57 -> 0.61Improved
add generated imagined scenes to real detector training data3D object detector · CLEVR, 50 annotations, IoU .3mean average precision: 0.43 -> 0.44Improved
increase imagined object count and randomize object locations, rather than relying on unique shape-style compositions3D object detector augmentation · CARLA and CLEVRmean average precisionImproved

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