vision-summary

module
v0.0.0-...-dc48539 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Oct 7, 2025 License: Apache-2.0

README

vision-summary

This is a Viam module containing models that summarize information from underlying vision models.

Getting started

First, create a machine in Viam.

Then, add a module from the Viam Registry and select the viam:vision-summary:count-classifier or viam:vision-summary:count-sensor model from the vision-summary module.

Configuration

viam:vision-summary:count-classifier

To configure the count-classifier model, use the following template:

{
  "detector_name": <string>,
  "chosen_labels": {
    <label1>: <float>,
    <label2>: <float>
  }
}
Attributes
Name Type Required? Default Description
detector_name string Required Name of the vision service to use as input. Must output a classifier tensor.
chosen_labels string Required Map of label names and required confidence values (between 0 and 1) to count the label in the summary.
count_thresholds object Required Maps summary category names to an integer value representing the (inclusive) upper bound of the range spanned by the category. Supports any number of custom category names. Each upper bound must be a unique integer.

To configure count_thresholds, define at least one category. For instance, the following example defines categories spanning the following ranges:

  • none: 0 to 10
  • low: 11 to 20
  • medium: 21 to 50
  • high: 51 to 100
"count_thresholds": {
  "none": 10,
  "low": 20,
  "medium": 50,
  "high": 100
}

### viam:vision-summary:count-sensor

To configure the `count-sensor` model, use the following template:

{ "camera_name": , "detector_name": , "chosen_labels": { : , : }, "count_thresholds": { "high": , "none": , "low": , "medium": } }


#### Attributes

| Name | Type | Required? | Default | Description |
| ---- | ---- | --------- | --------| ------------ |
| `camera_name` | string | **Required** | | Camera name to use for video input. |
| `detector_name` | string | **Required** | | Name of the vision service to use as input. Must output a classifier tensor. |
| `chosen_labels` | string | **Required** | | Map of label names and required confidence values (between 0 and 1) to count the label in the summary. |
| `poll_frequency_hz` | object | Optional | | How many times to summarize per minute. |
| `count_thresholds` | object | **Required** | | Maps summary category names to an integer value representing the (inclusive) upper bound of the range spanned by the category. Supports any number of custom category names. Each upper bound must be a unique integer. |

To configure `count_thresholds`, define at least one category. For instance, the following example defines categories spanning the following ranges:

- `none`: 0 to 10
- `low`: 11 to 20
- `medium`: 21 to 50
- `high`: 51 to 100

```json
"count_thresholds": {
  "none": 10,
  "low": 20,
  "medium": 50,
  "high": 100
}

Directories

Path Synopsis
cmd
module command
Package main is a module which serves the countclassifier vision service
Package main is a module which serves the countclassifier vision service

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL