Pith. sign in

REVIEW 1 cited by

CLIP-VAD: Exploiting Vision-Language Models for Voice Activity Detection

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.14509 v1 pith:SS3QO6AG submitted 2024-10-18 cs.CV

CLIP-VAD: Exploiting Vision-Language Models for Voice Activity Detection

classification cs.CV
keywords modelsvisualactivityapproachaudio-visualautomaticallyclipdata
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Voice Activity Detection (VAD) is the process of automatically determining whether a person is speaking and identifying the timing of their speech in an audiovisual data. Traditionally, this task has been tackled by processing either audio signals or visual data, or by combining both modalities through fusion or joint learning. In our study, drawing inspiration from recent advancements in visual-language models, we introduce a novel approach leveraging Contrastive Language-Image Pretraining (CLIP) models. The CLIP visual encoder analyzes video segments composed of the upper body of an individual, while the text encoder handles textual descriptions automatically generated through prompt engineering. Subsequently, embeddings from these encoders are fused through a deep neural network to perform VAD. Our experimental analysis across three VAD benchmarks showcases the superior performance of our method compared to existing visual VAD approaches. Notably, our approach outperforms several audio-visual methods despite its simplicity, and without requiring pre-training on extensive audio-visual datasets.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. SincQDR-VAD: A Noise-Robust Voice Activity Detection Framework Leveraging Learnable Filters and Ranking-Aware Optimization

    cs.SD 2025-08 conditional novelty 4.0

    A lightweight VAD model with a learnable sinc filterbank and a squared-margin ranking loss reports higher AUROC and F2 scores on AVA-Speech and ACAM using only 8.0k parameters.