AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

There Is More Than Meets the Eye: Self-Supervised Multi-Object Detection and Tracking With Sound by Distilling Multimodal Knowledge

Francisco Rivera Valverde, Juana Valeria Hurtado, Abhinav Valada

CVPR 2021 · Detection and segmentation › Open-set and few-shot detection · approach: Open-set and unknown object detection · 78 citations (OpenAlex)

Check 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
Distill from RGB, depth, and thermal teachers using MTA rather than a single RGB teacher with Ranking lossAudio student on MAVD · MAVD testmAP@ Avg: 44.05 -> 61.62Improved
Use multi-teacher MTA rather than averaging teacher activationsMM-DistillNet audio student · MAVD testmAP@ Avg: 51.63 -> 61.62Improved
Use MTA rather than Ranking loss for single-teacher distillationAudio student with RGB teacher · MAVD testmAP@ Avg: 44.05 -> 44.58Improved
Use MTA rather than Ranking loss for single-teacher distillationAudio student with depth teacher · MAVD testmAP@ Avg: 45.85 -> 55.41Improved
Add thermal teacher to RGB and depth teachers (ablation)MM-DistillNet audio student · MAVD testmAP@ Avg: 42.89 -> 61.10Improved
Add depth teacher to RGB and thermal teachers (ablation)MM-DistillNet audio student · MAVD testmAP@ Avg: 55.81 -> 61.10Improved
Add car-counting pretext-task initialization to the audio student (ablation)MM-DistillNet with three teachers · MAVD testmAP@ Avg: 61.10 -> 61.62Improved
Initialize the audio student with car-counting pretext-task weights (ablation)MM-DistillNet with three teachers · MAVD trainingtraining loss: 27.55% lowerImproved
Use Pairwise loss instead of MTA lossAudio student with RGB teacher · MAVD testmAP@ Avg: 40.45 vs 44.58Hurt
Average Ranking-loss teacher activations instead of integrating them with MTAAudio student with RGB, depth, and thermal teachers · MAVD testmAP@ Avg: 56.16 vs 61.62Hurt
Average AFD-loss teacher activations instead of integrating them with MTAAudio student with RGB, depth, and thermal teachers · MAVD testmAP@ Avg: 58.50 vs 61.62Hurt
Use MTA with r=4, t=9 instead of r=2, t=9 (ablation)Audio student with RGB teacher · MAVD testmAP@ Avg: 69.72 -> 60.13Hurt

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