The objective of this project is to investigate the images using Keras, TensorFlow, and Deep Learning techniques.
- You may resize data (i.e. into 128×128) as it would be more convenient.
- Use 80% of samples (positive and negative) for training and the remaining 20% for testing per category.
AssignmentTutorOnline
Question 1: Implement and select DCNN model that provides a good accuracy for dataset.
For designing the best DCNN model, consider the following criteria:
- Good initialization method (Xavier or He initializer)
- Activation function: ReLU
Question 2: Evaluated the performance of the model for different optimization functions:
- SGD
- Adam
- RMSProp
Question 3: Show and explain the structure of the model used (can be in word with flowcharts or on paper)
Question 4: Design a graphical user interface (GUI) that allows the user to select one image from a local file, run the testing process, and return the output as a message that says “You have chosen” either ‘POSITIVE ‘or ‘NEGATIVE’.
NOTE: The language used should be Python and to be executed in Anaconda. And use local files rather than using Drive module.