Learning Positive-Incentive Point Sampling in Neural Implicit Fields for Object Pose Estimation
This work addresses object pose estimation for robotics and computer vision, particularly in challenging scenarios, and is incremental as it builds on existing neural implicit field methods with novel sampling and equivariance enhancements.
The paper tackles the challenge of predicting canonical coordinates for unobserved camera-space regions in neural implicit fields for object pose estimation, which often leads to high uncertainty and inaccurate estimations. The proposed method, combining an SO(3)-equivariant convolutional implicit network and a positive-incentive point sampling strategy, outperforms state-of-the-art on three datasets, with significant improvements in scenarios like unseen poses, high occlusion, novel geometry, and severe noise.
Learning neural implicit fields of 3D shapes is a rapidly emerging field that enables shape representation at arbitrary resolutions. Due to the flexibility, neural implicit fields have succeeded in many research areas, including shape reconstruction, novel view image synthesis, and more recently, object pose estimation. Neural implicit fields enable learning dense correspondences between the camera space and the object's canonical space-including unobserved regions in camera space-significantly boosting object pose estimation performance in challenging scenarios like highly occluded objects and novel shapes. Despite progress, predicting canonical coordinates for unobserved camera-space regions remains challenging due to the lack of direct observational signals. This necessitates heavy reliance on the model's generalization ability, resulting in high uncertainty. Consequently, densely sampling points across the entire camera space may yield inaccurate estimations that hinder the learning process and compromise performance. To alleviate this problem, we propose a method combining an SO(3)-equivariant convolutional implicit network and a positive-incentive point sampling (PIPS) strategy. The SO(3)-equivariant convolutional implicit network estimates point-level attributes with SO(3)-equivariance at arbitrary query locations, demonstrating superior performance compared to most existing baselines. The PIPS strategy dynamically determines sampling locations based on the input, thereby boosting the network's accuracy and training efficiency. Our method outperforms the state-of-the-art on three pose estimation datasets. Notably, it demonstrates significant improvements in challenging scenarios, such as objects captured with unseen pose, high occlusion, novel geometry, and severe noise.