AutoSNAP: Automatically Learning Neural Architectures for Instrument Pose Estimation

Abstract

Despite recent successes, the advances in Deep Learning have not yet been fully translated to Computer Assisted Intervention (CAI) problems such as pose estimation of surgical instruments. Currently, neural architectures for classification and segmentation tasks are adopted ignoring significant discrepancies between CAI and these tasks. We propose an automatic framework (AutoSNAP) for instrument pose estimation problems, which discovers and learns the architectures for neural networks. We introduce 1)~an efficient testing environment for pose estimation, 2)~a powerful architecture representation based on novel Symbolic Neural Architecture Patterns (SNAPs), and 3)~an optimization of the architecture using an efficient search scheme. Using AutoSNAP, we discover an improved architecture (SNAPNet) which outperforms both the hand-engineered i3PosNet and the state-of-the-art architecture search method DARTS.

arXiv commentAccepted at MICCAI 2020 Preparing code for release at https://github.com/MECLabTUDA/AutoSNAP

References in corpus (8)

• VoxelMorph: A Learning Framework for Deformable Medical Image Registration
• Neural Architecture Search: A Survey
• DARTS: Differentiable Architecture Search
• Surgical Data Science: Enabling Next-Generation Surgery
• Neural Architecture Optimization
• CAI4CAI: The Rise of Contextual Artificial Intelligence in Computer Assisted Interventions
• Scalable Neural Architecture Search for 3D Medical Image Segmentation
• i3PosNet: Instrument Pose Estimation from X-Ray in temporal bone surgery