The pystar-fc from ykotseruba

computing peripheral map

I'm setting up pySTAR-FC and I noticed that PeripheralAttentionMap.salMap and PeripheralAttentionMap.periphMap` look exactly the same. In the code, it looks like the saliency map is blurred, but the blurred map is never used again:

pySTAR-FC/src/PeripheralAttentionalMap.py

Lines 43 to 47 in 9968ea0

    
           def computePeriphMap(self, mask): 
        
               blurredPeriphMap = cv2.GaussianBlur(self.salMap,(11,11),0) 
        
               self.periphMap = self.salMap.copy() 
        
               if mask: 
        
                   self.periphMap[self.periphMask==0] = 0

Is this a bug or am I missing something?

Best,
Matthias

error when using SALICONtf model and images batch

Hi Yulia,

Thanks for your open-source amazing work.

Since test.ini takes in one single image, I changed it to "batch" in test.ini to see the SALICONtf results in a batch of images. I got an unexpected error after the first image of the directory finished running. But if I use the images batch with "DeepGazeII" model as in other .ini files, the program runs well. Could you help?

Regards,
Qingyuan

Traceback (most recent call last):
  File "src/STAR_FC.py", line 47, in <module>
    main(sys.argv)
  File "src/STAR_FC.py", line 44, in main
    controller.run()
  File "/home/.../pySTAR-FC/src/Controller.py", line 68, in run
    self.setup(imgPath)
  File "/home/.../pySTAR-FC/src/Controller.py", line 56, in setup
    self.centralMap = CentralAttentionalMap(self.env.height, self.env.width, self.settings)
  File "/home/.../pySTAR-FC/src/CentralAttentionalMap.py", line 22, in __init__
    self.buSal = SALICONtf(weights='contrib/SALICONtf/models/model_lr0.01_loss_crossentropy.h5')
  File "/home/.../pySTAR-FC/src/SALICONtf.py", line 46, in __init__
    self.build_salicon_model()
  File "/home/.../pySTAR-FC/src/SALICONtf.py", line 152, in build_salicon_model
    fine_stream_input, fine_stream_output = self.build_vgg16(input_shape=(600, 800, 3), stream_type='fine')
  File "/home/.../pySTAR-FC/src/SALICONtf.py", line 140, in build_vgg16
    model.load_weights(os.path.dirname(os.path.abspath(__file__))+'/../contrib/SALICONtf/models/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5', by_name=False)
  File "/home/.../pySTAR-FC/py35-venv/lib/python3.5/site-packages/tensorflow/python/keras/engine/network.py", line 1516, in load_weights
    saving.load_weights_from_hdf5_group(f, self.layers)
  File "/home/.../pySTAR-FC/py35-venv/lib/python3.5/site-packages/tensorflow/python/keras/engine/saving.py", line 820, in load_weights_from_hdf5_group
    K.batch_set_value(weight_value_tuples)
  File "/home/.../pySTAR-FC/py35-venv/lib/python3.5/site-packages/tensorflow/python/keras/backend.py", line 2880, in batch_set_value
    get_session().run(assign_ops, feed_dict=feed_dict)
  File "/home/.../pySTAR-FC/py35-venv/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 929, in run
    run_metadata_ptr)
  File "/home/.../pySTAR-FC/py35-venv/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1095, in _run
    'Cannot interpret feed_dict key as Tensor: ' + e.args[0])
TypeError: Cannot interpret feed_dict key as Tensor: Tensor Tensor("Placeholder_7:0", shape=(3, 3, 64, 128), dtype=float32) is not an element of this graph.

ykotseruba / pystar-fc Goto Github PK

pystar-fc's People

Contributors

Stargazers

Watchers

Forkers

pystar-fc's Issues

computing peripheral map

error when using SALICONtf model and images batch

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent

	def computePeriphMap(self, mask):
	blurredPeriphMap = cv2.GaussianBlur(self.salMap,(11,11),0)
	self.periphMap = self.salMap.copy()
	if mask:
	self.periphMap[self.periphMask==0] = 0