Synthetic Particle Generator

This is a python program for generating simulated images of spherical or elliptical particles. Particle information is generated randomly from given statistical parameters, and this information is stored in a SQLite database. The resulting labeled dataset can be used to train models to perform general edge detection tasks.

Getting Started

1. Pull this repository
2. Activate the virtual environment by running $ pipenv shell from the project directory
3. Create settings file at /config/mysettings.yaml or use default file /config/default.yaml
4. Interact through cli.py entrypoint
   1. Initialize sqlite database file and create tables python cli.py dbinit
   2. Generate 5 samples with the given settings: python cli.py simulate ./config/mysettings.yaml 5

Configuration

The parameters used for generating each image are randomly sampled based on statistics provided in the settings.yaml file.

	name: Name of this parameter set for notekeeping purposes
	xdim: X dimension of the images to be produced
	ydim: Y dimension of the image to be produced
	mean_r: Average particle radius
	sd_r: Standard deviation of particle radius
	mean_aspect: Average particle aspect ratio
	sd_aspect: Standard deviation of aspect ratio
	mean_n: Average number of particles plotted in each image
	sd_n: Standard deviation of number of particles
	mean_theta: Average angle (in radians) of an occlusion
	sd_theta: Standard deviation of occlusions
	n_dist: A list of tuples (n, p) where n is the number of occlusions and p is the probability
	gutters: A list (gutterx, guttery) that provides the padding on the edge of the image, where particles won't be placed
    min_dist: Factor determining the minimum distance between particle centers. 0 to allow overlapping particles, 1 is approximately 1 radius of padding so particles will not quite overlap