Real-time semantic segmentation on FPGAs for autonomous vehicles with hls4ml
This addresses the need for efficient, low-latency hardware acceleration in autonomous driving systems, though it is incremental as it builds on existing compression and quantization methods.
The paper tackled real-time semantic segmentation for autonomous vehicles by deploying a compressed ENet CNN on FPGAs, achieving a latency of 4.9 ms per image with less than 30% resource usage and reducing it to 3 ms for batch processing while maintaining accuracy on Cityscapes.
In this paper, we investigate how field programmable gate arrays can serve as hardware accelerators for real-time semantic segmentation tasks relevant for autonomous driving. Considering compressed versions of the ENet convolutional neural network architecture, we demonstrate a fully-on-chip deployment with a latency of 4.9 ms per image, using less than 30% of the available resources on a Xilinx ZCU102 evaluation board. The latency is reduced to 3 ms per image when increasing the batch size to ten, corresponding to the use case where the autonomous vehicle receives inputs from multiple cameras simultaneously. We show, through aggressive filter reduction and heterogeneous quantization-aware training, and an optimized implementation of convolutional layers, that the power consumption and resource utilization can be significantly reduced while maintaining accuracy on the Cityscapes dataset.