Papers
1
Total Citations
10
H-Index
1
About
Meng Du is a leading researcher in energy-efficient voice-control and speech-processing hardware, with a focus on keyword spotting (KWS) and speaker verification systems. Their most-cited work, "14.8 KASP: A 96.8% 10-Keyword Accuracy and 1.68μJ/Classification Keyword Spotting and Speaker Verification Processor Using Adaptive Beamforming and Progressive Wake-Up" (2024, 10 citations), introduces a groundbreaking processor that achieves 96.8% accuracy on 10-keyword tasks while consuming only 1.68μJ per classification. This design addresses critical challenges in smart home, intelligent robot, and wearable applications, particularly the sensitivity of existing KWS processors to human-voice noise from nearby individuals. By integrating adaptive beamforming and a progressive wake-up mechanism, Du’s processor significantly improves robustness and energy efficiency, setting a new benchmark for always-on voice interfaces. This work exemplifies Du’s contributions to low-power, high-accuracy speech processing, enabling practical deployment in battery-constrained devices. Their research bridges the gap between algorithmic innovation and hardware implementation, making voice control more reliable and accessible in noisy real-world environments.
Research Focus
Key Achievements
Top Papers
- 1