Creating a Classification Application

34 STEPS

1. First, select the Application Builder on the left side of the screen. You may notice many different options, but for now we will select Classification under Vision AI Applications.

2. Type the Name of the Classification application and Press Enter. When naming the application, it is important to name it something meaningful that directly relates to the application purpose.

3. Click Next blue button, and you will proceed to the Input Subjects page.

4. An Input Subject is simply a group or collection of images/ videos which will be then used by the Application for Training. The Input subject is the media which has not yet been processed by the Classification Application.

5. Type an Input Subjects name. Again please consider a name that will help you differentiate the Input Subject in the future.

6. Click Create a new subject in the current scenario, named as animals 

7. Click Next blue button in the end of the page, so you can proceed to the Output Subjects page

8. The Output Subject of the Classification application is where you will find the labeled and sorted images. Usually the Output Subject gets populated with media (images/ videos) after successful Training of the Application. 

9. Type Output Subjects name. Again please consider the name, as it should be directly related with the Subject Of Interest displayed in the image.

10. Click Create a new subject in the current case scenario, named as dogs

11. Type the second Output Subjects name, as the Classification app requires at least 2 and up to 2,200 Subjects to detect inside the images.

12. Click Create a new subject in our case named as cats

13. Click Next blue button

14. The yellow message banner informs you that instead of 2 subjects Classification application, you could choose a Detection application, which gives more control over precision and recall. Click Next

15. In this final step you will select some of the advanced settings. The Frames Per Second Requirement is related with the Application Model and the images which will be processed. Click the Frames per Second Requirement to change the value.  

16. Type Frames per Second Requirement in a numeric value.

17. Click Complete which will close the following window and will open the Application Summary window for you.

18. Click the Input Subject which is always displayed in the beginning of the Application pipeline. In our case it is called 'animals'. 

19. Click Upload Media to add the initial Dataset to to the Input Subjects. Once it has been uploaded to the Input Subject, the Media goes for training by the User.

20. Click select files or select folder to upload the initial Dataset of images/ videos to the Input Subjects of the Application

21. Click Next once you have selected the files or the folder containing the Media

22. Click Upload 

23. Click on the Application Name in our case it is animals-classifier 

24. Click Provide Feedback, so you can start training the Application Model

25. Select the option (Subject of Interest) displayed in the image on the left side and Click Confirm button

26. Click Replays, so you can provide more feedback, since the Training is a continuous process of uploading images and Providing Feedback. 

27. Select the Input Subject. In current scenario select animals as previously we named it 'animals'.

28. With your mouse, click and drag the Probability Filter, and drop it at the beginning of the scale

28b. Drop

29. Scroll down and select the Force Feedback option which will enforce the Feedback, and will populate the Feedback queue.

30. Click Replay Subjects blue button in the end of the window

31. Click Confirm once you have selected the Subject Of Interest (in the current scenario it is dogs)

32. Click Confirm once you have selected the Subject Of Interest 

33. When there are no images left in the Feedback Queue No pending feedback screen will be displayed. Click Quit

34. That's it. You're done.