---
title: "SegmentationModel"
sidebar_label: "SegmentationModel"
description: "The clika_runtime.modelverse.models SegmentationModel class."
---

<!-- Generated by tools/api_reference/generate_api_docs.py. Do not edit. -->

A loaded segmentation model (panoptic or instance).

## `labels` (property)



## `segment`

```python
segment(self, image: 'clika_runtime.Tensor', *, score_threshold: 'float | None' = None, mask_threshold: 'float | None' = None) -> 'Any'
```

The segmentation of an [H, W, 3] image: a ``SegmentationMap`` with the
UInt8 [H, W] id mask at the image's own size (``background_id`` where
nothing is painted) and one ``SegmentationSegment`` (segment_id,
label_id, label, score) per painted id. The thresholds default to the family's
own decode; a family whose decode keeps fixed thresholds refuses a
set one.