• Rien d’enregistré
Terminé!
Concepteur: James Studt (2)
Prix: Gratuit
Classements: 0 
Commentaires: 0 Publier un commentaire
Listes: 0 + 0
Points: 0 + 0 ¡
+1 Points

Description

Image ML Annotator
Nouvel  Image ML Annotator helps you build labeled image datasets for object detection on macOS.

If you are training a Create ML object detector, the slow part is often not the training itself. It is getting a clean set of images, drawing boxes around the objects you care about, labeling them consistently, and exporting everything in the format your training workflow expects. Image ML Annotator is a focused desktop tool for that job.

Start by creating a document and importing a folder of images. The app keeps your images, label names, detector choice, and annotations together in one package, so you can stop and resume work later without reconstructing your setup.

Next, define the labels you want to use. For a simple project, that might be one label. For a multi-class detector, you can configure several labels and switch between them quickly from the keyboard.

Then move through your images and mark what matters. Draw rectangles around the objects you want to teach your model to recognize, adjust them by dragging or resizing, and assign labels with fast shortcuts. If an image contains none of your target objects, mark that clearly so your dataset reflects reality instead of leaving ambiguous gaps.

To speed things up, you can also import an existing Core ML detector. Image ML Annotator can use that model to propose rectangles and labels for the current image. You review the proposal, accept it, correct it, resize it, move it, relabel it, or delete it. This makes the app especially useful when you already have a first-generation model and want to use it to help create the next round of training data.

As you work, the app is designed for iterative labeling rather than one-shot export. You can import more images later, paste in screenshots, remove bad images, revisit older annotations, and keep refining the document over time.

When you are ready to train, export your annotated images and Create ML annotations​.json into a new folder. If you already have an existing training export, you can also use Export Into to merge the current work into a larger corpus without manually reconciling files and annotations. That makes it practical to grow a dataset in stages instead of rebuilding it from scratch each time.

A typical workflow looks like this:

1. Import a folder of images.
2. Configure the labels for the objects you care about.
3. Optionally import a previously trained detector to generate proposals.
4. Review each image, drawing or correcting boxes and assigning labels.
5. Mark images with no target objects when appropriate.
6. Export the finished annotations for Create ML.
7. Train a model.
8. Bring that model back into Image ML Annotator to accelerate the next labeling pass.

Image ML Annotator is built for people who need a practical way to turn piles of images into usable object detection training data, especially when that work happens in repeated cycles of label, train, improve, and label again.
plus ↓
Ad

Captures d'écran

#1. Image ML Annotator (macOS) De: James Studt
#2. Image ML Annotator (macOS) De: James Studt

Historique des prix

  • Prix actuel : Gratuit
  • Minimum: Gratuit
  • Maximum: Gratuit
Suivre l'évolution des prix

Informations supplémentaires

Nouveautés

  • Version: 1.0
  • Mis à jour:

Concepteur

Listes (0)

0 ☹️
  • Image ML Annotator

Points (0)

0 ☹️

Classements (0)

0 ☹️

Commentaires (0)

Soyez le premier à commenter 🌟
«Image ML Annotator». Plateforme : macOS. Catégorie : Développeurs. Développeur : «James Studt». Première version : . Dernière mise à jour : . Prix actuel : gratuit. Ce titre n'a pas encore reçu de notes ou d'avis sur AppAgg. Langues disponibles : English. AppAgg suit l'historique des prix, les notes et les commentaires des utilisateurs pour «Image ML Annotator». Suivez les futures réductions et mises à jour via RSS. AppAgg n'héberge pas d'applications et ne distribue pas de logiciels. Toutes les marques, logos et captures d'écran appartiennent à leurs propriétaires respectifs.
Image ML AnnotatorImage ML Annotator URL courte: Copié!
  • 🌟 Partager
  • Mac App Store

Similaires

Découvrez également

Vus récemment

Opérateurs de recherche compatibles avec AppAgg
Ajouter à AppAgg
AppAgg
Lancez-vous, c'est gratuit.
Inscription
Connexion