CHiME-7

The CHiME-7 dataset/benchmark contains multi-channel recordings of conversational speech and is used to evaluate distant automatic speech recognition (ASR) and diarization systems.

Papers using CHiME-7 (10)

Efficient and Generalizable Speaker Diarization via Structured Pruning of Self-Supervised Models2025 · 13 citesRecent Trends in Distant Conversational Speech Recognition: A Review of CHiME-7 and 8 DASR Challenges2025 · 1 citesThe CHiME-7 Challenge: System Description and Performance of NeMo Team's DASR System2023 · 5 citesApplying LLMs for Rescoring N-best ASR Hypotheses of Casual Conversations: Effects of Domain Adaptation and Context Carry-over2024 · 2 citesNTT speaker diarization system for CHiME-7: multi-domain, multi-microphone End-to-end and vector clustering diarization2023BUT CHiME-7 system description2023Multi-CMGAN+/+: Leveraging Multi-Objective Speech Quality Metric Prediction for Speech Enhancement2023Automatic channel selection and spatial feature integration for multi-channel speech recognition across various array topologies2023Remixed2Remixed: Domain adaptation for speech enhancement by Noise2Noise learning with Remixing2023The CHiME-8 DASR Challenge for Generalizable and Array Agnostic Distant Automatic Speech Recognition and Diarization2024