AMD64 | ARM64 |
---|---|
A simple docker based application to store and label images. Metadata of images will automatically be extracted and stored as well. The api is build with FastAPI and uses a sqlite database to store the data. Images will be saved in the filesystem. The purpose of the project is to collect and label images for a machine learning model. To train the model, you can simply request an image from the api. You can see the available endpoints below.
There is also a custom frontend to manage the dataset, update labels and metadata. More information coming soon!
This application is designed to run in a docker environment. In order to use the image you can easily pull the latest image from docker hub. You can find the images for AMD and for ARM here. Use the following command to pull the docker on your local machine:
docker pull castox/image-dataset-api_amd64:latest
docker pull castox/image-dataset-api_arm64:latest
You can add labels to an image. The labels are saved as a list of strings. You can add as many labels as you want. Those can be used to search for images. You can also update the labels of an image.
You can save and request metadata from an image. If an image is sent to the api, all available metadata will be extracted and saved. You can also update the metadata of an image. The metadata is saved as a dict and can be parsed into a json object.