# Toward a Procedural Fruit Tree Rendering Framework for Image Analysis

## TL;DR

This paper introduces a procedural fruit tree rendering framework using Blender and Python to generate labeled datasets for training deep learning models in image analysis, especially for robotic fruit harvesting, with domain randomization features.

## Contribution

It presents a novel, flexible framework for generating synthetic, labeled datasets with domain randomization to improve deep learning training in fruit image analysis.

## Key findings

- Generated datasets facilitate training of deep learning models.
- Framework supports domain randomization for robustness.
- Accelerates dataset creation for robotic fruit harvesting.

## Abstract

We propose a procedural fruit tree rendering framework, based on Blender and Python scripts allowing to generate quickly labeled dataset (i.e. including ground truth semantic segmentation). It is designed to train image analysis deep learning methods (e.g. in a robotic fruit harvesting context), where real labeled training datasets are usually scarce and existing synthetic ones are too specialized. Moreover, the framework includes the possibility to introduce parametrized variations in the model (e.g. lightning conditions, background), producing a dataset with embedded Domain Randomization aspect.