tags

Tags are used to provide information about a certain data file, which may be specified in the tags entry in the JSON file.

How to use?

In VisionAI format, the semantic segmentation RLE format are using tags information to indicate every pixel belonging class.

When your data not included segmentation, it is not necessary to provide tags information.

Schema

Example

"tags": {
    "7fbac044-c160-11ed-afa1-0242ac120002": {
        "ontology_uid": "",
        "type": "semantic_segmentation_RLE",
        "tag_data": {
            "vec": [{
                "type": "values",
                "val": ["background", "car", "sky", "vegetation"],
                "name": ""
            }]
        }
    }
}

The RLE format value in the VisionAI format is a compressed representation of the pixel-wise information in an image. For example, if the mask pixels from left to right are "car, car, car, sky, sky, sky, sky", the corresponding RLE value will be

#3V1#4V2

In this example, the numbers "3" and "4" represent the counts of consecutive pixels for the "car" and "sky" classes, respectively. The class numbers are indicated by "V1" and "V2", which are defined in the values.

tags {}

name	description	type	required
${TAG_UUID}	The id of the tag. It uses UUID32 as a key.	object	true
ontology_uid	The field of ontology id. For the semantic_segmentation_RLE case, it is a required item and should be left empty.	string	true
type	In the VisionAI format, the field refers to the type of information being provided. This field is only available for the value of "semantic_segmentation_RLE" and is used to indicate the class number information for the semantic segmentation binary RLE value.	string	true
tag_data	The semantic segmentation RLE information about this type.	object	true

tag_data {}

name	decription	type	required
vec	The array type of information. In this case, it contains the class mapping array in RLE.	object	true
type	The value type of the information. Please fill in “values“ in this case.	string	true
val	This value is used to map class names to values in order to describe the semantic segmentation RLE format. For example, a list of class names such as ["background", "car", "sky", "vegetation"] would mean that "background" is represented as V0 and "car" is represented as V1 in the RLE format.	array	true
name	It is a required field, please leave it as ““.	string	true

---

Use Case

semantic segmentation

To describe a semantic segmentation dataset with one camera sensor:

• sensor: camera (#camera1)

• ontology

  • background

  • person

  • bicycle

  • car

  • motorcycle

  • airplane

  • bus

  • train

  • truck

  • boat

  • trafficlight

Example Code

pagesemantic segmetation