Unity-Technologies / Indoor-Pet-Detection

A demo project for indoor pet detection using Unity generated synthetic data.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Indoor Pet Detection Example

License Unity

This project provides scripts necessary to train a model for dog detection in an indoor home environment, with Unity generated Synthetic Data, and evaluate your model. For this example, we will use synthetic data to train our model and real data from COCO and OIDSv6 to fine tune and test our model.

We also show how we can upload our own custom assets, and change randomizer parameters to change the data as required. Unity Computer Vision Datasets (UCVD) is the cloud platform used for generating large scale datasets.

Checkout our example at - Demo

Instructions

  1. Prerequisites
  2. Create a synthetic dataset for dog detection with UCVD
  3. Training, Fine-tuning & Evaluation
  4. Results

Learning More About Unity Synthetic Data

  1. Unity Computer Vision Datasets
  2. Overview on how UCVD works
  3. Unity Perception
  4. Create your own assets on UCVD

Sample Datasets

Results

Pre-training Fine tune(real data) AP Data generation time(mins)
ImageNet + no synthetic 120051.16---
ImageNet + 5k synthetic 120057.1833
ImageNet + 10k synthetic 120058.4235
ImageNet + 40k synthetic 120063.0350
ImageNet + 100k synthetic 120065.484
Scratch + no synthetic 120023.16---
Scratch + 5k synthetic 120036.1633
Scratch + 10k synthetic 120036.9735
Scratch + 40k synthetic 120047.350
Scratch + 100k synthetic 120053.7884

About

A demo project for indoor pet detection using Unity generated synthetic data.

License:Other


Languages

Language:Jupyter Notebook 99.9%Language:Python 0.1%Language:Shell 0.0%