StandardSim

StandardSim is a large-scale photorealistic synthetic dataset for retail environments with over 25k examples. Modalities include rgb, depth, before/after, instance segmentation labels, depth, and normals.

Project Page: https://standard-ai.github.io/Standard-Sim/

ICIAP 2022 Paper: https://arxiv.org/abs/2202.02418

Data: https://drive.google.com/drive/folders/1KRjuL1wxe2VqYKKIOAKa0-FXjTNgr4TX

Code for training models coming soon.

Change Detection Task

Change Detection Labels and Model Predictions

Data Modalities

Dataset Comparison

Shield:

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

This license is inclusive of both this codebase and the StandardSim data release.