Classifier based Image Sorter

This app can use an image classifier to detect any number of target labels, then sort those images into 1 of 2 folders.

Requirements

Sign up for alwaysAI
Install alwaysAI tooling

Worflow Options

This app can run on either a desktop/laptop development machine or on an edge device such as a Jetson Nano, Dragonboard, or Raspberry Pi that has docker installed. See the above install instruction link for more details.

Configuration

This app makes use of key-values in the config.json file for set up and variable assignments:

Key	Value Type	Description
found_folder	string	Optional location within the app's root folder to put images that have target objects in them. Default is `output_images/found`
empty_folder	string	Optional location within the app's root folder to put images that had no detected target objects. Default is `output_images/not_found`
source_folder	string	Optional location within the app's root folder to process images from. Default is `source_images`
classifier	object	A dictionary with classifier information. See below for expected key-values

Classifier Information:

Key	Value Type	Description
model_id	string	The model id as found from alwaysAI's Model Catalog. Note that there should be a duplicate of this string value in the higher level 'models' key-value pair which is used by alwaysAI's CLI tool. This app has not been keyed to use that string as any number of models may be downloaded to an app for other purposes. To add or remove models for download, see the docs here.
confidence_level	float	This is the minimum required confidence level value required by a detected object to be considered found. Each model has a confidence_level value that's provided with each classification prediction. Many models will use a float value from 0.0-1.0 with 1.0 = 100%, but some models may have a different value based system, like `squeezenet_v1_1`
target_labels	array of strings	List of all labels this model should be filtering / looking for. This will usually be a subset of all the available labels this model is capable of detecting. If no labels are specified nothing will be returned, in other words the `found_folder` will be empty after processing

Running

Use the alwaysAI CLI to build and start this app:

Configure (once): aai app configure

Build: aai app install

Run: aai app start

Support

Docs: https://dashboard.alwaysai.co/docs/getting_started/introduction.html

Community Discord: https://discord.gg/alwaysai