This article demonstrates the classification of a publicly available image using IBM Watson Visual Recognition.
What you'll need:
Skill level: Easy
Time required: 3 minutes
- Scheduler by Quickwork as a trigger--for triggering the journey
- IBM Watson Visual Recognition: Action
Configuring a trigger
- Under the Trigger section, choose the Scheduler by Quickwork app from the drop-down menu in the App field.
- Select the trigger event, New scheduled event, from the drop-down menu in the Trigger Event field.
- Set the Interval for One day as we need to fetch the breaking news on a daily basis.
- Set the date and time of your choice in the Start At field and keep the Custom Payload field empty:
Configuring IBM Watson Visual Recognition action
- Under the Steps section, choose the IBM Watson Visual Recognition service from the drop-down menu in the App field.
- Select the Classify image action from the drop-down menu in the Action field.
- Click the Link an account button to establish a new connection with IBM Watson Visual Recognition service. You can also connect an existing account if you have one.
- The Image URL input field will open. Here, specify the publicly available image URL that you want to classify with each aspect. The size of the image should not be more than 10 MB.
E.g., https://s3-ap-southeast-2.amazonaws.com/media1.mydeal.com.au/blog/post/blog_image_20180219221810890.jpg:
Executing the journey
Save the changes in a journey and click the Start Journey button. You'll be taken to the History tab automatically. Click on the journey ID that just got executed. In the Steps section, click the IBM Watson Visual Recognition action bar to expand and switch to the Output tab to see the output:
You'll see a list of components and class in which the image has been classified. After analyzing more precisely, it also returns with a dominant class of the image URL specified. The specified example image URL contained happy and cheerful people with red filtration effect, and so is the result.
Comments
0 comments
Please sign in to leave a comment.