Winning Edge AI: Automatic Annotation & Data Augmentation for Image Classification
How do you push the limits of image classification on ultra-low-power devices?
In this presentation, Anas Benalla, Data Scientist at Assayin, shares his innovative approach to the Edge AI Wake Vision Challenge that earned him 1st place on the final leaderboard.
💡 Learn how he combined:
⚡ Automatic annotation using an efficient object detection model
🎛️ Data augmentation techniques like random eraser, rotation, zoom, and color shift
🖥️ Optimization for ultra-low-power microcontrollers
Watch as Anas explains how he reached 84.5% accuracy and outlines additional methods for improving AI model performance at the edge.
📌 If you’re working on edge-deployed vision systems, this talk is packed with practical strategies and inspiration for your own projects.
🎥 Watch now and explore the future of Edge AI:
👉 https://youtu.be/NGmr2IChJ9g
#EdgeAI #AIChallenge #WakeVisionChallenge #ImageClassification #DataAugmentation #AutomaticAnnotation #Assayin #Microcontrollers #AIOptimization #AIOnDevice #TinyML #TechForGood #AIInnovation
source
