AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

A Light Weight Model for Active Speaker Detection

Junhua Liao, Haihan Duan, Kanghui Feng, Wanbing Zhao, Yanbing Yang, Liangyin Chen

CVPR 2023 · Architectures and efficiency › RNNs and normalization · approach: Dataset condensation and flow estimation · 56 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
Use kernel sizes 3 and 5 together instead of kernel size 3 alone (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 93.0 -> 94.1Improved
Increase kernel size from 5 to 7 (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 93.4 -> 93.4No clear effect
Replace the lightweight visual encoder with the TalkNet visual encoder (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 94.1 -> 92.8Hurt
Replace split 2D and 1D visual convolutions with 3D convolution (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 94.1 -> 92.9Hurt
Replace the lightweight audio encoder with ResNet-18 (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 94.1 -> 93.4Hurt
Use 2D convolutions instead of split 1D audio convolutions (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 94.1 -> 93.6Hurt
Add temporal modeling with forward GRU instead of direct FC prediction (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 88.0 -> 92.6Improved
Use bidirectional GRU instead of forward GRU (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 92.6 -> 94.1Improved
Replace forward GRU with a transformer detector (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 92.6 -> 91.5Hurt
Add the proposed loss with visual auxiliary supervision (ablation)Our Method · AVA-ActiveSpeaker validationmAP (%): 93.1 -> 94.1Improved

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