Indoor Scene Analysis & Classification

There are many things that can be recognized in the interior which may help in creating amazing applications

The Wizart Vision Analysis API includes a set of different computer vision solutions based on neural networks.

Camera data - provides a set of camera parameters that correspond to the moment of shooting for the uploaded image. These are parameters such as FoV (focal length), camera height and camera angles that allow you to build a virtual camera that meets all the requirements for working with this image in 3D or AR modes. A virtual camera is usually used to obtain projection matrices and perform correct ray tracing in a virtual scene based on a specific photo.

Interior type - determines whether the interior belongs to one of the described types. List of types that can be determined:

Class
0
1
2
3
4
5
6
7
8
9
10
11
Type
living room
kitchen
bathroom
utility room
dining room
renovation
hall
bedroom
studio room
children's room
home office
office

Image quality - api endpoint contains information about the quality of shooting, about noise and blur artifacts at photo.

see documentation
Response example:
        
{   
    "camera": {
        "fov": 48.38069074577039,
        "pitch": 0.020306486046132832,
        "roll": 0.001955529658362931,
        "height": 1.5
    },
    "interior_type": "living room",
    "image_info": {
        "bad_target_confidence": 0.8,
        "blurry_confidence": 0.8,
        "dark_confidence": 0.8,
        "noisy_confidence": 0.8:
    }
}
        
    

Learn more

Indoor Scene Semantic Segmentation
Object Detection In Indoor Scenes
3D Room Layout Reconstruction
Technologies for Home Interior

Ready to get started?

Create your free API token