SegmentInfo

Represents COCO-style per-segment metadata for one segment in a panoptic or instance segmentation. Fields include id, category_id, and optionally area, bbox, and iscrowd. This datatype is used as the schema for each element in the segments_info list of PanopticSegmentationAnnotation within the Telekinesis ecosystem.

Field

Required

Field	Type	Description
id	int	Segment identifier (int).
category_id	int	Category id for this segment (int).

Optional

Field	Type	Description
area	float	Segment area. Default 0.0.
bbox	list[float] or np.ndarray	Bounding box [x, y, width, height]; length 4, dtype float. Default [0.0, 0.0, 0.0, 0.0].
iscrowd	int	Whether the segment is a crowd region. Default 0.

Methods

Method	Description
to_dict()	Returns a dictionary with keys: id, category_id, area, bbox, iscrowd.

Example

From the datatypes examples:

import numpy as np
from datatypes import datatypes
from loguru import logger

# ------------------------------------------------
# 1. Create SegmentInfo instance
# ------------------------------------------------
segment_info = datatypes.SegmentInfo(
    id=1,
    category_id=1,
    area=100.0,
    bbox=np.array([0.0, 0.0, 10.0, 10.0]),
    iscrowd=0,
)

# ------------------------------------------------
# 2. Access data via to_dict
# ------------------------------------------------
segmentation_info = segment_info.to_dict()
logger.info("SegmentInfo id={}", segmentation_info["id"])
logger.info("SegmentInfo category_id={}", segmentation_info["category_id"])
logger.info("SegmentInfo area={}", segmentation_info["area"])
logger.info("SegmentInfo bbox={}", segmentation_info["bbox"])
logger.info("SegmentInfo iscrowd={}", segmentation_info["iscrowd"])