Probability Maps

[bondruy]

Thank you Sir for the wonderful code I have a question，when I test on MIT300， how do I get the the log density predictions of a probability model

[alexanderkroner]

Hey, what you can do to generate log density maps is to first retrieve the output from the network and then convert it as follows:

import tensorflow as tf

graph_def = tf.GraphDef()

with tf.gfile.Open("model_salicon_gpu.pb", "rb") as file:
    graph_def.ParseFromString(file.read())

input_plhd = tf.placeholder(tf.float32, (None, None, None, 3))

[predicted_maps] = tf.import_graph_def(graph_def,
                                       input_map={"input": input_plhd},
                                       return_elements=["output:0"])

sum_per_image = tf.reduce_sum(predicted_maps[0], keep_dims=True)
log_density_maps = tf.log(tf.divide(predicted_maps[0], sum_per_image) + 1e-8)

with tf.Session() as sess:
    saliency = sess.run(log_density_maps, feed_dict={input_plhd: input_img})

You just need to download a pre-trained model from here and feed an input image of your choice.

[bondruy]

Thank you very much