Mikyong Ji

Papers

Total Citations

H-Index

About

Mikyong Ji is a researcher whose work lies at the intersection of speech processing, human-robot interaction, and ubiquitous computing. Her primary research focus is on developing robust speaker identification systems for real-world, distant-talking environments—a critical challenge for enabling natural and reliable communication between humans and autonomous robots. Ji’s major contributions center on the innovative use of multiple microphones to overcome the degradation in speech signals caused by distance and noise. She pioneered a "soft channel selection" method, which intelligently weights and combines identification results from an array of microphones, rather than relying on a single, potentially unreliable channel. This approach, detailed in her most-cited work "Text-Independent Speaker Identification using Soft Channel Selection in Home Robot Environments" (20 citations), significantly improves classification accuracy in dynamic settings like robot companions. Her earlier foundational work, "Reliable Speaker Identification Using Multiple Microphones in Ubiquitous Robot Companion Environment" (8 citations), established the framework for this multi-microphone strategy. By addressing the practical limitations of far-field speech recognition, Ji’s research directly enables more intuitive and effective human-robot collaboration, making her a key contributor to the advancement of socially aware robotics.