Lightweight image segmentation

Abstract

Deploying advanced image segmentation tasks on mobile devices struggle with the demands of sophisticated deep learning models. Image segmentation, a critical task in computer vision, has seen remarkable advancements through deep learning. However, the implementation of these computationally intensive models on mobile devices is hindered by their large size and resource demands. The project aims to develop a mobile-friendly, lightweight deep learning architecture for image segmentation, drawing inspiration from DeepLabV3’s capabilities. The goal is to balance the trade-off between accuracy and speed, thereby making advanced image segmentation feasible on mobile platforms.