BlankSkip: Early-exit Object Detection onboard Nano-drones
This addresses the problem of enabling real-time object detection for nano-drones, which is incremental as it adapts an existing early-exit approach to a dense task.
The paper tackles the challenge of deploying object detection on nano-drones with tight computational constraints by proposing BlankSkip, an adaptive network that uses early exits for frames with no objects, achieving up to 24% average throughput improvement with only a 0.015 mAP drop compared to a static detector.
Deploying tiny computer vision Deep Neural Networks (DNNs) on-board nano-sized drones is key for achieving autonomy, but is complicated by the extremely tight constraints of their computational platforms (approximately 10 MiB memory, 1 W power budget). Early-exit adaptive DNNs that dial down the computational effort for "easy-to-process" input frames represent a promising way to reduce the average inference latency. However, while this approach is extensively studied for classification, its application to dense tasks like object detection (OD) is not straightforward. In this paper, we propose BlankSkip, an adaptive network for on-device OD that leverages a simple auxiliary classification task for early exit, i.e., identifying frames with no objects of interest. With experiments using a real-world nano-drone platform, the Bitcraze Crazyflie 2.1, we achieve up to 24% average throughput improvement with a limited 0.015 mean Average Precision (mAP) drop compared to a static MobileNet-SSD detector, on a state-of-the-art nano-drones OD dataset.