Comprehensive survey of CNN hardware acceleration across FPGA, ASIC, and VLSI implementations spanning five real-world domains. Presents unified benchmarks, cross-domain optimisation strategies, and key research challenges for efficient, low-power, real-time AI systems; manuscript under review.
asic fpga deep-learning vlsi hardware-software-co-design real-time-inference cnn-hardware-acceleration
-
Updated
Apr 3, 2026