Beautification application

Use this application to beautify the selfie by improving the skin condition or adding makeup. The app has three different modes.

Attach an application to your dataset:

The Beautification application can be attached to the dataset, just like the Face Application.

#Face Beautification
BEAUTIFICATION_APPLICATION_ID = "c879e501-b203-47fa-b333-d4877f0eef99" 

resp =
        "dataset_id": dataset_id,
        "algorithms_flags": [
                "name": "beautification_config",
                "values": ["healthy_colored", "healthy", "light_healthy"],
    headers={"Authorization": f"Bearer {token}"},


    "Beautification application has been successfully attached to dataset:",

The field algorithms_flags should contain a values list that describes the degree of beautification effect. The image can be beautified with only one mode"values": [ "healthy_colored" ]as well as with two or three modes "values": [ "healthy", "healthy_colored" ]. By default "values":["healthy"]

NB! Please note that once you have attached an application to a dataset with the configuration described in algorithms_flags, you cannot change them; if you would like to configure the application differently, you will need to create a new dataset and attach it to the application with another algorithms_flags.

Upload an image to your dataset

The application supports only frontal facial images, i.e., you should specify side_id == 1 during the uploading process

Beautification options

The app has the following three beautification options:

  1. Light healthy: apply this mode to slightly improve facial skin quality, reduce skin pigmentation, and decrease scarring (see light_healthy in the illustration below).

  2. Healthy: apply this mode to increase skin evenness, reduce redness, and decrease skin irritation. Facial skin will look visibly younger (see healthy in the illustration below).

  3. Healthy colored: apply this mode to combine the “Healthy” mode with lightly colored makeup (see healthy_colored in the illustration below).


The results can be parsed from aux_image_type from an API response

                  "id": 46,
                  "name": "Light healthy",
                  "tech_name": "beautification_original_light_healthy"
                  "id": 47,
                  "name": "Healthy",
                  "tech_name": "beautification_original_healthy"
                  "id": 47,
                  "name": "Healthy colored",
                  "tech_name": "beautification_original_healthy_colored"

Last updated