Zhongnan Qu

TL;DR: This work proposes to develop pedestrian detectors that unlock the potential of the event data by leveraging multi-cue information and different fusion strategies, and introduces three different event-stream encoding methods based on Frequency, Surface of Active Event (SAE) and Leaky Integrate-and-Fire (LIF).

TL;DR: This contribution proposes an online multi-target tracking system utilizing for neuromorphic vision sensors, which is the first neuromorph vision system in intelligent transportation systems, and integrates an online tracking-by-clustering system running at a high frame rate, which far exceeds the real-time capabilities of traditional frame based industry cameras.

TL;DR: Adaptive Loss-aware Quantization (ALQ), a new MBN quantization pipeline that is able to achieve an average bitwidth below one-bit without notable loss in inference accuracy, is proposed.

TL;DR: This work addresses the challenges of bringing Machine Learning to MCUs, where it focuses on the ubiquitous ARM Cortex-M architecture and proposes an implementation-aware design as a cost-effective method for verification and benchmarking.

TL;DR: A deep neural network for grasping detection is developed that considers the angle learning problem as classification instead of regression, and performs high detection accuracy on the Event-Grasping dataset with 93% precision at an object-wise level split.