Visual Recognition Challenge

The IBM Watson Visual Recognition service is a powerful AI tool that identifies image content. The service comes with the following pretrained models, but can also be customized to recognize custom classes.

  • General: A trained model that lets you use an active and large data set to gain insights from your own pictures.
  • Food: Similar to the general model, but uses a specific food data set. You can use the model to find any type of food; one of the main applications is for the catering and restaurant industry, especially if you want to create a specific menu.
  • Explicit: Enables you to analyze whether an image contains inappropriate content.
  • Custom: Lets you create a custom model with your own pictures and train it to get better results.

In this Visual Recognition Challenge for the Digital Developer Conference, you’ll complete a hands-on lab. You’ll use food images, Watson Studio, and Watson Visual Recognition to identify food, and use a model that utilizes a specialized vocabulary of over 2000 foods to identify meals, food items, and dishes with enhanced accuracy.

Learning objectives

After completing this tutorial you will be able to:

  • Create a Watson Visual Recognition service
  • Score the identified objects

Prerequisites

This tutorial can be completed using an IBM Cloud Lite account.

  1. Create an IBM Cloud account.
  2. Log into IBM Cloud.

Estimated time

You can complete this task in no more than 15 minutes.

Hands-on lab overview

The outline below provides a high-level overview of the steps included in the lab instructions.

Step 1: Watson Studio

We will create a Watson Studio service, create a Project and Watson Visual Recognition model to identify images in several classes.

Watson Studio setup

Create a Cloud Object Storage instance

  1. Go to the IBM Cloud Catalog.
  2. Search for Object in the IBM Cloud Catalog.
  3. Click the Object Storage service tile.

    Cloud Object Storage Catalog screenshot

  4. Click Create. The Cloud Object Storage instance is created.

    Cloud Object Storage Catalog screenshot

Create a Watson Studio service instance

  1. Create a Watson Studio service instance from the IBM Cloud Catalog.
  2. Search for Studio in the IBM Cloud Catalog.

    Watson Studio Catalog screenshot

  3. Click the Watson Studio service tile.

    Watson Studio Service screenshot

  4. Click Create. The Watson Studio service is created.

  5. Click Get Started.

    Watson Studio Launch screenshot

  6. Log in with your IBM Cloud account.

  7. Work through the introductory tutorial to learn about Watson Studio.

    Watson Studio Welcome screenshot

Watson Studio projects

Projects act as your workspace where you can organize your resources, such as data, collaborators, and analytic tools like notebooks and models.

Create a New Project

  1. Click Create a Project.
  2. Click the Create an empty project tile.

    Watson Studio New project screenshot

  3. Enter a name for your project: VRChallenge

  4. The Cloud Object Storage instance you created in a previous step should be prefilled. Click Create.

    Watson Studio New project screenshot

Create a new Visual Recognition Service

  1. Click the Settings tab in your project page.
  2. In the Associated Services section, click Add service and then select Watson. A list of Watson services opens.

    New Watson service screen

  3. In the Visual Recognition Service card, click Add.

  4. Select the Lite plan and note the features.
  5. Scroll to the bottom and click Create. Ensure that “Dallas” is entered as the region and then click Confirm.
  6. In the IBM Cloud window, open the IBM Cloud menu and click Resource List.
  7. Click the Visual Recognition service to open the Visual Recognition Service page and then click Launch Watson Studio to launch the service dashboard.

    VR service page screenshot

Step 2: Testing

In this section of the lab, we’ll use the Food model, one of the pre-built Watson Visual Recognition models.

  1. In the Food model panel, click Test.

    VR Test screenshot

  2. Click the Test tab. Load the test images into the Test page by copying the following two images — test1.jpeg and test2.jpeg — to your desktop and then clicking and dragging them into the Test page.

    Test images
    test image 1 test image2

    VR Test screenshot

  3. Inspect the scores returned by the Watson Visual Recognition Food Classifier.

Summary

To get more hands-on experience using Watson Visual Recognition through step-by-step tutorials, deep-dive videos, and complete examples of working code, check out the Watson Visual Recognition learning path on IBM Developer.

Hania Ibrahim