<p>Contactless human-machine interaction (HMI) is rapidly evolving, yet it remains constrained by the latency, redundancy, and power consumption inherent in conventional frame-based vision sensors. While bio-inspired event-driven sensors offer a low-power alternative, existing architectures are often complex or fail to accurately encode the magnitude of light intensity changes. Herein, we report a solution-processed, two-terminal event-driven sensory device based on CuIn(Se,S)₂ colloidal quantum dots (QDs) integrated with an Sb-doped TiO₂ layer. Unlike traditional dynamic vision sensors (DVS), this device exhibits a transient photoresponse that encodes both the polarity and the magnitude of light intensity variations into the output current amplitude. This preservation of magnitude information significantly enhances the feature extraction capability, leading to faster convergence and superior clustering performance in gesture recognition. Based on these unique optoelectronic properties, we constructed a hierarchical HMI system that synergizes the strengths of event-based and frame-based sensing. The system utilizes the event-driven sensor for low-latency gesture control of an unmanned aerial vehicle (UAV) and a frame-based sensor for high-precision gaze control of an unmanned ground vehicle (UGV). The proposed system achieves a gesture recognition accuracy of more than 92.5% while substantially reducing data redundancy, offering a promising strategy for efficient, robust, and low-cost intelligent interaction systems.</p>