🎉 TakuNet Paper Accepted at WACVW 2025: Breaking the Speed Barrier in Drone Vision

Created on January 20, 2025

2025

:newspaper: New Publication! Our paper “TakuNet: an Energy-Efficient CNN for Real-Time Inference on Embedded UAV systems in Emergency Response Scenarios” has been accepted at IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW) 2025!

The work presents a lightweight CNN architecture optimized for real-time emergency response classification on battery-powered drones, achieving state-of-the-art efficiency on NVIDIA Jetson and Raspberry Pi.

Authors: Daniel Rossi, Guido Borghi, Roberto Vezzani
Conference: WACVW 2025, Tucson, Arizona
Code: github.com/danielrossi1/TakuNet