Adding Image Recognition to the App

Watson Visual Recognition

Image recognition capability is powered by Watson Visual recognition service on IBM Cloud. Create a Watson Visual Recognition instance on IBM Cloud. For more information, see here.

Once configured, you can now create a new Mobel and add classes to it. You can drag and drop images in to the Builder and then train your Model on those images. Once the training is complete, you can either download the CoreML model or use the Model in a AI control in your app.

To enable a Visual Recognition in your app, perform the following steps:

1. Click Watson and then click Image Recognition. This displays the Work with Watson Visual Recognition screen.

   

2. Click Connect to your Watson Visual Recognition instance.

   

3. Enter the API key details and specify the URL of your Watson Visual Recognition instance.

4. Provide a Name to your Image Recognition instance in the app and click Connect. This displays the dashboard for your model.

   

5. Click Add new model to create a new model. This will display Create a new Model popup.

   

6. Enter the Model name and click Create. This will display the classes for that model and a Negative class.

   

7. Click Add new class. This will display a popup to specify a name for the new class.

   

8. Enter the Class name for the new class and click Create. This will display the workspace to add your images for training the model.

   

9. Add the images to the model either by drag and drop them into the workspace or use Browse to access the images.

10. You can go back to your workspace after adding the images and test by clicking Test Model.

    

11. In the Try your model section, add an image and then result is displayed.