# Stacked 1D convolutional networks for end-to-end small footprint voice trigger detection

Source: [Stacked 1D convolutional networks for end-to-end small footprint voice trigger detection](<https://www.sota2.com/research/paper/68fa68151e209131cea1a0c3>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T22:29:00.055Z

Authors: Takuya Higuchi, Mohammad Ghasemzadeh, Kisun You, Chandra Dhir

Published: 2020-08-08T00:32:55.000Z

## Abstract

We propose a stacked 1D convolutional neural network (S1DCNN) for end-to-end small footprint voice trigger detection in a streaming scenario. Voice trigger detection is an important speech application, with which users can activate their devices by simply saying a keyword or phrase. Due to privacy and latency reasons, a voice trigger detection system should run on an always-on processor on device. Therefore, having small memory and compute cost is crucial for a voice trigger detection system. Recently, singular value decomposition filters (SVDFs) has been used for end-to-end voice trigger detection. The SVDFs approximate a fully-connected layer with a low rank approximation, which reduces the number of model parameters. In this work, we propose S1DCNN as an alternative approach for end-to-end small-footprint voice trigger detection. An S1DCNN layer consists of a 1D convolution layer followed by a depth-wise 1D convolution layer. We show that the SVDF can be expressed as a special case of the S1DCNN layer. Experimental results show that the S1DCNN achieve 19.0% relative false reject ratio (FRR) reduction with a similar model size and a similar time delay compared to the SVDF. By using longer time delays, the S1DCNN further improve the FRR up to 12.2% relative.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Keyword Spotting on 2000 hours (test)](<https://www.sota2.com/research/sota/keyword-spotting-on-2000-hours-test.md>): Keyword Spotting; 2000 hours (test); FRR: 1.18.

- [Wake-word detection on Wake-word detection (test)](<https://www.sota2.com/research/sota/wake-word-detection-on-wake-word-detection-test.md>): Wake-word detection; Wake-word detection (test); FRR: 1.99.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2008.03405>)

[PDF](<https://arxiv.org/pdf/2008.03405.pdf>)