noise-image-generator

Simple tool to generate noise images that can be used for testing image-based AI or anything else.

https://github.com/systemsourcer/noise-image-generator

Science Score: 44.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
    Found CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
  • Academic publication links
  • Academic email domains
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (8.7%) to scientific vocabulary
Last synced: 6 months ago · JSON representation ·

Repository

Simple tool to generate noise images that can be used for testing image-based AI or anything else.

Basic Info
  • Host: GitHub
  • Owner: SystemSourcer
  • License: mit
  • Default Branch: main
  • Size: 5.86 KB
Statistics
  • Stars: 0
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created 10 months ago · Last pushed 10 months ago
Metadata Files
Readme License Citation

README.md

Noise-image-generator

Simple tool to generate noise images that can be used for testing image-based AI or anything else.

Background

There are some Advantages of Testing an image-absed AI with with Noise-Only Images:

  • Classification Separation Assessment: If the model makes reliable predictions on noise images, it could indicate that it is not effectively separating classes and may yield false positives. This draws attention to possible weaknesses in the model architecture or training dataset.

  • Model Robustness: Testing with images that consist only of noise can help determine how robust the model is against non-informative data. A well-trained model should be able to distinguish between noise and actual objects. If the model makes incorrect predictions on such images, it may indicate potential overfitting or inadequate generalization.

  • Threshold Development: Such tests can help in defining thresholds and confidence criteria. If a model performs well in detecting objects but fails to make reliable predictions with noise-only images, it can help in determining when an object should be regarded as detected.

  • Evaluation of Preprocessing Techniques: You can also assess how well preprocessing and augmentation techniques support the model.

Features

The Code generates and saves monochrome, RGB, multi and hyperpsectral images including the following features:

  • Setting width, height and nummber of channels / bands.
  • Setting number of images to generate
  • Setting destination for saing the images
  • Saving the images as numpy arraytogether with the resulting PNG / TIFF file

Installation

  1. Clone the repository or just donload the nig.py file.
  2. Execute the nig.py file in a terminal: ```bash python nig.py

Owner

  • Login: SystemSourcer
  • Kind: user

Citation (CITATION.cff)

cff-version: 1.2.0
message: "If you use this software, please cite it as below."
author:
- family-names: "Mielke"
  given-names: "Simon"
  orcid: "https://orcid.org/0009-0008-6490-2720"
title: "Noise Image Generator"
date-released: 2025-05-07
url: "https://github.com/SystemSourcer/Noise-image-generator/"

GitHub Events

Total
  • Push event: 5
  • Pull request event: 2
Last Year
  • Push event: 5
  • Pull request event: 2