Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SW-ASR: A Context-Aware Hybrid ASR Pipeline for Robust Single Word Speech Recognition

Manali Sharma, Riya Naik, Buvaneshwari G

TL;DR

This work tackles robust single-word automatic speech recognition (SW-ASR) in noisy and bandwidth-constrained environments. It introduces a modular pipeline that denoises and normalizes audio, uses a hybrid front end combining Whisper and Vosk for initial transcripts, and employs a versatile verification layer with multiple matching strategies. The verification component, including context-guided and LLM-based matching, yields the largest accuracy gains on real-world, low-quality data while maintaining real-time latency. The approach demonstrates practical viability for telephony and messaging platforms, enabling reliable single-word interpretation in open- and closed-vocabulary settings.

Abstract

Single-word Automatic Speech Recognition (ASR) is a challenging task due to the lack of linguistic context and sensitivity to noise, pronunciation variation, and channel artifacts, especially in low-resource, communication-critical domains such as healthcare and emergency response. This paper reviews recent deep learning approaches and proposes a modular framework for robust single-word detection. The system combines denoising and normalization with a hybrid ASR front end (Whisper + Vosk) and a verification layer designed to handle out-of-vocabulary words and degraded audio. The verification layer supports multiple matching strategies, including embedding similarity, edit distance, and LLM-based matching with optional contextual guidance. We evaluate the framework on the Google Speech Commands dataset and a curated real-world dataset collected from telephony and messaging platforms under bandwidth-limited conditions. Results show that while the hybrid ASR front end performs well on clean audio, the verification layer significantly improves accuracy on noisy and compressed channels. Context-guided and LLM-based matching yield the largest gains, demonstrating that lightweight verification and context mechanisms can substantially improve single-word ASR robustness without sacrificing latency required for real-time telephony applications.

SW-ASR: A Context-Aware Hybrid ASR Pipeline for Robust Single Word Speech Recognition

TL;DR

This work tackles robust single-word automatic speech recognition (SW-ASR) in noisy and bandwidth-constrained environments. It introduces a modular pipeline that denoises and normalizes audio, uses a hybrid front end combining Whisper and Vosk for initial transcripts, and employs a versatile verification layer with multiple matching strategies. The verification component, including context-guided and LLM-based matching, yields the largest accuracy gains on real-world, low-quality data while maintaining real-time latency. The approach demonstrates practical viability for telephony and messaging platforms, enabling reliable single-word interpretation in open- and closed-vocabulary settings.

Abstract

Single-word Automatic Speech Recognition (ASR) is a challenging task due to the lack of linguistic context and sensitivity to noise, pronunciation variation, and channel artifacts, especially in low-resource, communication-critical domains such as healthcare and emergency response. This paper reviews recent deep learning approaches and proposes a modular framework for robust single-word detection. The system combines denoising and normalization with a hybrid ASR front end (Whisper + Vosk) and a verification layer designed to handle out-of-vocabulary words and degraded audio. The verification layer supports multiple matching strategies, including embedding similarity, edit distance, and LLM-based matching with optional contextual guidance. We evaluate the framework on the Google Speech Commands dataset and a curated real-world dataset collected from telephony and messaging platforms under bandwidth-limited conditions. Results show that while the hybrid ASR front end performs well on clean audio, the verification layer significantly improves accuracy on noisy and compressed channels. Context-guided and LLM-based matching yield the largest gains, demonstrating that lightweight verification and context mechanisms can substantially improve single-word ASR robustness without sacrificing latency required for real-time telephony applications.
Paper Structure (13 sections, 2 figures, 2 tables)

This paper contains 13 sections, 2 figures, 2 tables.

Table of Contents

  1. Introduction and Related Work
  2. Experimental Setup
  3. Data
  4. Data Curation
  5. System Model and Design
  6. Initial Audio Transcription
  7. Similarity Matching
  8. Cosine Similarity
  9. Levenshtein Distance
  10. LLM Matching
  11. Context Guided Matching
  12. Results and Discussion
  13. Conclusion

Figures (2)

  • Figure 1: SW-ASR: (1) The single-word speech audio processed to improve quality. (2)Hybrid: Whisper & Vosk used to generate initial transcript. (3) Candidate matching performed using three approaches. (4) Contextual Concatenation for Generalization.
  • Figure 2: Average time taken across datasets for each approach