Don't Listen to Me: A Lightweight, Low-Latency Model for Own-Voice Cancellation in Far-Field Speech Enhancement

[system]

---

This is a companion discussion topic for the original entry at https://arxiv.org/abs/2606.23332