The given TfRecord record_test.tfrecord is generated by using TfRecord_Generator.py.
The dataset consist of three images.
Modify and complete the file task_run_cnn.py to solve the upcomming three tasks:
Complete the method CNN.getInput() of the class CNN CNN to read the individuall images of the dataset record_test.tfrecord.
Define a 5x5 Gausian Filter and apply it in the method CNN.inference(img) to the input img.
Use a Tensorflow graph for this taks. CNN.inference(img) should return an Tensor.
Finally, apply the Gausian-Filter to the image of Homer Simpson within a Tensorflow session. Generate a Github branch and provide the final image of Homer and a snapshot of the Gausian-Filter. Use a Tensorflow Saver object for this task.