Skip to content

Quick Start

This guide walks you from a fresh EXACT installation to your first annotated image.

1. Start EXACT

git clone https://github.com/DeepMicroscopy/Exact.git
cd Exact
cp exact/exact/settings.py.example exact/exact/settings.py
docker compose -f docker-compose.yml up -d --build

Open http://localhost:8000/ and log in with exact / exact.

2. Create a Team

A Team is the top-level organizational unit. All ImageSets and their annotation types belong to a team.

  1. On the home page, click New Team in the left sidebar.
  2. Enter a team name and confirm.

3. Create an ImageSet

An ImageSet is a collection of related images (e.g., one study, one slide batch).

  1. Click New imageset inside your team.
  2. Enter a name and click the checkmark.
  3. Open the imageset by clicking its card.

4. Upload Images

Inside the imageset view:

  1. Click Upload images.
  2. Drag and drop images onto the upload area, or use the file picker.

Supported formats include JPEG, PNG, BMP, and all formats supported by OpenSlide (SVS, NDPI, SCN, MRXS, …), as well as NIfTI (.nii, .nii.gz).

After upload, images are processed and thumbnails are generated automatically.

5. Define Annotation Types

Before annotating you need at least one Annotation Type, which defines the shape and label of an annotation.

  1. From the home page, scroll to the bottom and open Administration.
  2. Create a Product (a logical grouping of annotation types for a study).
  3. Inside the product, click Add Annotation Type.
  4. Choose a name, shape (bounding box, circle, polygon, …), and color.
  5. Assign the product to your imageset via Edit imageset → Products.

6. Annotate

  1. Click any image thumbnail to open the annotator.
  2. Select an annotation type from the left toolbar.
  3. Draw the annotation on the image:
    • Bounding box: click and drag
    • Circle: click center, drag to set radius
    • Polygon: click to place vertices, double-click to close
  4. Press Enter or click Save to confirm. Press Escape to cancel.

7. Verify

Verification is a second-pass quality check. An annotation is considered verified once a second annotator marks it correct.

  1. Open the Verification view from the imageset page.
  2. Step through annotations and press ✓ (correct) or ✗ (wrong).

8. Export

  1. In the imageset view, open Export.
  2. Choose or create an export format (CSV, JSON, COCO, …).
  3. Download the resulting file.

Next Steps