# SG-DOR: Learning Scene Graphs with Direction-Conditioned Occlusion Reasoning for Pepper Plants

arXiv: https://arxiv.org/abs/2603.06512 · HTML: https://mlsanity.com/paper/2603.06512/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 13 of its 27 references that are in the map.
- First posted on arXiv 6 Mar 2026.

## Builds on (most cited first, 13 of 13)

- [PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space](https://mlsanity.com/paper/1706.02413/) (2017)
- [Strategies for Pre-training Graph Neural Networks](https://mlsanity.com/paper/1905.12265/) (2019)
- [3D Scene Graph: A Structure for Unified Semantics, 3D Space, and Camera](https://mlsanity.com/paper/1910.02527/) (2019)
- [Occluded Video Instance Segmentation: A Benchmark](https://mlsanity.com/paper/2102.01558/) (2021)
- [Panoptic Mapping with Fruit Completion and Pose Estimation for Horticultural Robots](https://mlsanity.com/paper/2303.08923/) (2023)
- [VISO-Grasp: Vision-Language Informed Spatial Object-centric 6-DoF Active View Planning and Grasping in Clutter and Invisibility](https://mlsanity.com/paper/2503.12609/) (2025)
- [Attention-driven Next-best-view Planning for Efficient Reconstruction of Plants and Targeted Plant Parts](https://mlsanity.com/paper/2206.10274/) (2022)
- [Towards Scene Understanding with Detailed 3D Object Representations](https://mlsanity.com/paper/1411.5935/) (2014)
- [NBV-SC: Next Best View Planning based on Shape Completion for Fruit Mapping and Reconstruction](https://mlsanity.com/paper/2209.15376/) (2022)
- [Safe Leaf Manipulation for Accurate Shape and Pose Estimation of Occluded Fruits](https://mlsanity.com/paper/2409.17389/) (2024)
- [GO-VMP: Global Optimization for View Motion Planning in Fruit Mapping](https://mlsanity.com/paper/2503.03912/) (2025)
- [HortiBot: An Adaptive Multi-Arm System for Robotic Horticulture of Sweet Peppers](https://mlsanity.com/paper/2403.15306/) (2024)
- [Efficient Manipulation-Enhanced Semantic Mapping With Uncertainty-Informed Action Selection](https://mlsanity.com/paper/2506.02286/) (2025)

## Abstract (arXiv)

Robotic harvesting in dense crop canopies requires effective interventions that depend not only on geometry, but also on explicit, direction-conditioned relations identifying which organs obstruct a target fruit. We present SG-DOR (Scene Graphs with Direction-Conditioned Occlusion Reasoning), a relational framework that, given instance-segmented organ point clouds, infers a scene graph encoding physical attachments and direction-conditioned occlusion. We introduce an occlusion ranking task for retrieving and ranking candidate leaves for a target fruit and approach direction, and propose a direction-aware graph neural architecture with per-fruit leaf-set attention and union-level aggregation. Experiments on a multi-plant synthetic pepper dataset show improved occlusion prediction (F1=0.73, NDCG@3=0.85) and attachment inference (edge F1=0.83) over strong ablations, yielding a structured relational signal for downstream intervention planning.