Supervisely
Last updated
Was this helpful?
Last updated
Was this helpful?
Easily import your videos with annotations in the Supervisely format. The Supervisely json-based annotation format supports such figures: rectangle, line (polyline), polygon, point, bitmap (mask), graph (keypoints). It is a universal format that supports various types of annotations and is used in the Supervisely platform.
Enterprise users have access to "Import as links" option, which supports import of this format with annotations. This option might be beneficial in many cases, as it allows data import to Supervisely platform without re-uploading, maintaining a single source and speeding up import process.
To step up import speed even further you can compress all annotation files (.json's) into an archive and import it together with the images. (Note: This method is format-dependent and may not apply to all formats.)
Supported video formats: .avi, .mp4, .3gp, .flv, .webm, .wmv, .mov, and .mkv
With annotations: yes
Supported annotation format: .json.
Data structure: Information is provided below.
Example data: .
Both directory and archive are supported.
Recommended directory structure:
Struggled with the structure? No worries! All videos will be uploaded to a single dataset, so you don't have to worry about the full project structure in Supervisely format. All you need is to prepare videos with annotations and meta.json file (recommended).
Items even can be placed in any subdirectories or the root directory. Just make sure that the annotation file names match the video file names (e.g. annotaion file video_1.jpg.json is for the video video_1.jpg) and that the annotation file format is correct (we will provide an example in the next section). The application will do the rest.
For each video, we store the annotations in a separate json file named video_name.video_format.json with the following file structure:
Fields definitions:
size - string - is equal to image(frame) size
description - string - (optional) - this field is used to store the text we want to assign to the video. In the labeling intrface it corresponds to the 'data' filed.
key - string, unique key for a given video (used in key_id_map.json to get the video ID)
frames - list of frames of which the video consists. List contains only frames with an object from the 'objects' field
index - integer - number of the current frame
figures - integer - list of objects which the current frame contains
framesCount - integer - total number of frames in the video
objectKey - string - unique key for a given object (used in key_id_map.json)
labelerLogin - string - the name of a user who created the current figure
geometryType - "cuboid_3d" - class shape
geometry - a dictionary containing indicators of location, rotation and dimensions of cuboids
Fields definitions for objects field:
key - string, a unique key for the given object (used in key_id_map.json to get the object ID)
classTitle - string - the title of a class. It's used to identify the class shape from the meta.json file
tags - list of strings that will be interpreted as object tags
labelerLogin - string - the name of the user that added this figure to the project
Fields description for figures field:
key - string, a unique key for the given figure (used in key_id_map.json to get the figure ID)
objectKey - string, a unique key for the given object (used in key_id_map.json to get the object ID).
geometryType - "rectangle" -class shape
geometry - geometry of the object
classTitle - string - the title of a class. It's used to identify the class shape from the meta.json file
labelerLogin - string - the name of the user that added this figure to the current frame
Key_id_map.json file is optional. It is created when annotating the video inside Supervisely interface and sets the correspondence between the unique identifiers of the video, object and the frame on which the object is located. If you annotate manually, you do not need to create this file. This will not affect the work being done.
Json format of key_id_map.json:
Fields definitions:
objects - dictionary, where the key is a unique string, generated inside Supervisely environment to set correspondence of current object in annotation, and values are unique integer ID corresponding to the current object
figures - dictionary, where the key is a unique string, generated inside Supervisely environment to set correspondence of object on current frame in annotation, and values are unique integer ID corresponding to the current frame
videos - dictionary, where the key is unique string, generated inside Supervisely environment to set correspondence of video in annotation, and value is a unique integer ID corresponding to the current video
tags - dictionary, where the keys are unique strings, generated inside Supervisely environment to set correspondence of tag on current frame in annotation, and values are a unique integer ID corresponding to the current tag
Project meta file meta.json is recommended to be present in the project directory. It contains classes and tags definitions for the project. If it is not present, it will try to create it from the annotations. Learn more about the meta.json file .
tags - list of strings that will be interpreted as video
objects - list of
