A Cocktail-Party Benchmark: Multi-Modal dataset and Comparative Evaluation Results

TL;DR

This paper introduces a new multi-modal dataset and benchmark for recognizing and separating overlapping multi-party conversations using audio, visual, and contextual cues, advancing research in natural multi-party speech recognition.

Contribution

It presents the MCoRec task, a novel benchmark with a multi-modal dataset for multi-party conversation recognition in challenging overlapping scenarios, including baseline systems and evaluation results.

Findings

Visual cues significantly improve recognition accuracy.

Audio-only systems struggle with high speech overlap.

Multi-modal approaches outperform audio-only baselines.

Abstract

We introduce the task of Multi-Modal Context-Aware Recognition (MCoRec) in the ninth CHiME Challenge, which addresses the cocktail-party problem of overlapping conversations in a single-room setting using audio, visual, and contextual cues. MCoRec captures natural multi-party conversations where the recordings focus on unscripted, casual group chats, leading to extreme speech overlap of up to 100% and highly fragmented conversational turns. The task requires systems to answer the question "Who speaks when, what, and with whom?" by jointly transcribing each speaker's speech and clustering them into their respective conversations from audio-visual recordings. Audio-only baselines exceed 100% word error rate, whereas incorporating visual cues yields substantial 50% improvements, highlighting the importance of multi-modality. In this manuscript, we present the motivation behind the task, outline the data collection process, and report the baseline systems developed for the MCoRec.