# SegICP: Integrated Deep Semantic Segmentation and Pose Estimation

## TL;DR

SegICP is a novel integrated approach combining deep semantic segmentation and multi-hypothesis point cloud registration to achieve real-time, accurate 6-DOF pose estimation of objects, improving robotic perception in complex scenarios.

## Contribution

It introduces SegICP, which couples CNN-based segmentation with point cloud registration for robust, real-time object recognition and pose estimation without initial seeds.

## Key findings

- Achieves 1cm position error in real time.
- Attains less than 5° angle error.
- Outperforms previous methods on benchmark dataset.

## Abstract

Recent robotic manipulation competitions have highlighted that sophisticated robots still struggle to achieve fast and reliable perception of task-relevant objects in complex, realistic scenarios. To improve these systems' perceptive speed and robustness, we present SegICP, a novel integrated solution to object recognition and pose estimation. SegICP couples convolutional neural networks and multi-hypothesis point cloud registration to achieve both robust pixel-wise semantic segmentation as well as accurate and real-time 6-DOF pose estimation for relevant objects. Our architecture achieves 1cm position error and <5^\circ$ angle error in real time without an initial seed. We evaluate and benchmark SegICP against an annotated dataset generated by motion capture.