[ANSWER] BIM400 ASSIGNMENT 7: SIMPLE MACHINE LEARNING APPLICATION EXERCISE

Background

Image classification refers to the process by which an image is recognized and matched to a specific word.  For instance, when you see an image of a dog, you immediately classify it as an object that can be described with the word “dog.” Although computers can easily read text, they have more difficulty classifying images, usually because the same word can describe creatures that appear quite different, such as a Great Dane and a Chihuahua.

With machine learning, computers can be trained on a variety of dog images and gain competence in correctly classifying new ones. You will create a model to distinguish between different objects in this activity. Watch the tutorial on the teachable machines (link in the instructions) explaining what to do.

Instructions 

1. Create image samples for two different objects.

1. 1. Take at least four pictures each of examples of two different objects (e.g., cats, dogs, books, bowls, cups, pencils, hands, plants, etc.). These will be used for training.

1. 1. Take two more pictures of examples of the objects that are not the same as any of your training images.  These will be used to test your model.

1. Go to Teachable Machine. Choose Image Project and then Standard Image Model.

1. For Class 1, upload pictures of your first object without
   Note: Full answer to this question is available after purchase.
   including your testing image, and click next to Class 1 to rename the class to the object’s name. Repeat the process for Class 2, with your second set of pictures, and rename the class with the name of the object. Click on Train Model and wait for the process to complete.

1. Test your model.

1. 1. Click on Export Model.  The programming code at the bottom of the screen is your model.  The software saves only this code and not your actual images.

1. 1. Choose Upload Model. Copy the URL link for your model and enter it in your address bar so you can test it.

1. 1. Drag each of your testing images to the box, one at a time, and note the model’s classification as well as its confidence level.

Answer the following questions: 

1. How well did your model perform when testing your images?
2. How confident was the model for each image?
3. Choose two similar objects to classify, such as paperback and hardback books or domestic cats and bobcats, take photos for training and testing, and train a new model.   How well did the model perform? How could you improve the accuracy of the model under those circumstances?

Length: This assignment must be 1-2 pages (excluding the title and reference page).

References: Include the assignment-supplied resources.