Deep Learning for Computer Vision

Learning the facial key points

As discussed in the earlier topics, there are a few parameters that are to be defined while calculating the key facial points. We will use the following code to define these parameters:

image_size = 40no_landmark = 10no_gender_classes = 2no_smile_classes = 2no_glasses_classes = 2no_headpose_classes = 5batch_size = 100total_batches = 300

Next, allow a few placeholders for the various inputs. 

image_input = tf.placeholder(tf.float32, shape=[None, image_size, image_size])landmark_input = tf.placeholder(tf.float32, shape=[None, no_landmark])gender_input = tf.placeholder(tf.float32, shape=[None, no_gender_classes])smile_input = tf.placeholder(tf.float32, shape=[None, no_smile_classes])glasses_input = tf.placeholder(tf.float32 ...