Image detection

Detect objects in an image

Upload an image to detect objects, draw probability boxes, and return an annotated image with optional labels.

Image

Object detection boxes with probability

Cloud

Image

Min confidence

Line width

Show labels

Object detection boxes with probability result

Open in Fuser

Uses AI on Fuser’s servers. This is a saved example.

How tools work

A tool is a small app you describe in plain words. Try this one here, then keep it on your Fuser canvas and connect it to the rest of your work.

A tool on the Fuser canvas with its controls and result

1. Try it here

Add your own file or text and adjust the controls. Live tools run in your browser, so nothing is uploaded.

Two tools chained together on the Fuser canvas

2. Open it in Fuser

Sign up free and the tool opens on your canvas with this example wired in, ready to connect to images, video or text from other nodes.

A Tool node with a description of the tool to build

3. Describe your own

Need something different? Describe the tool you want and Fuser builds it, with the controls you asked for.

Ways to use this detector

Turn an input image into a boxed, optionally labeled result for visual review and follow-on image workflows.

Review detected objects

Inspect an image with boxes placed around detected objects and probability information.

Create visual callouts

Add object-detection boxes to photos when you need the detected areas clearly marked.

Filter uncertain detections

Raise the minimum confidence setting to focus the result on detections that meet your chosen threshold.

Make a labeled image

Keep Show labels enabled to include labels with the detection boxes in the returned image.

Simplify the annotation

Turn Show labels off when you want the boxes without visible object labels.

Prepare an image for another node

Pass the annotated result into a larger Fuser workflow that accepts an image as its next input.

Frequently asked questions

It takes one image and returns an image with boxes around detected objects. The tool is built to include probability information, and labels can be shown or hidden.

Provide one image in the required image input. The image is the source for the object detections and the returned annotated result.

Min confidence sets the threshold for which detections are included in the result. Increase it to filter out detections below your chosen level, or lower it to allow more detections through.

Line width changes how thick the detection boxes appear in the returned image. It can be adjusted from 1 to 10.

Yes. Use the Show labels switch to include or hide labels on the detection boxes. It is enabled by default.

Sign in to Fuser, provide your own image, adjust the controls, and run the tool. This is a Cloud tool, so each run uses credits on Fuser’s servers.

Yes. Use the returned image as an input for another compatible node, or connect this detector after an image-producing node. This lets the boxed result become part of a larger workflow.