How to have similar predictions as Places365 Demo

[northtree]

Please make sure that the boxes below are checked before you submit your issue.

Thank you!

• [x] Check that you are up-to-date with the master branch of Keras. You can update with:

  pip install git+git://github.com/keras-team/keras.git --upgrade --no-deps

• [x] If running on TensorFlow, check that you are up-to-date with the latest version. The installation instructions can be found here.

• [x] If running on Theano, check that you are up-to-date with the master branch of Theano. You can update with:

  pip install git+git://github.com/Theano/Theano.git --upgrade --no-deps

• [x] Provide a link to a GitHub Gist of a Python script that can reproduce your issue (or just copy the script here if it is short).

---

From the official demo website: http://places2.csail.mit.edu/demo.html It will return below results from test image: http://places2.csail.mit.edu/imgs/demo/6.jpg

How could we have the confidence from scene categories, such as food_court (0.690)

Predictions: Type of environment: indoor Scene categories: food_court (0.690), cafeteria (0.163) Scene attributes: no horizon, enclosed area, man-made, socializing, indoor lighting, cloth, congregating, eating, working

The current vgg16_places_365 just return categories without confidence (also return different ranking).

PREDICTED SCENE CATEGORIES: cafeteria food_court restaurant_patio banquet_hall restaurant

[northtree]

@GKalliatakis It seems that the official demo used ResNet instead of VGG16

[albertobeto]

Different network architecture and training will lead to different prediction score. But you can print the confidence with something like:

  predictions_to_return = 5
  preds = model.predict(image)[0]
  top_preds = np.argsort(preds)[::-1][0:predictions_to_return]
  top_preds_score = np.sort(preds)[::-1][0:predictions_to_return]

  for i in range(0, 5):
      print("{:.2f}".format(top_preds_score[i]*100)+'% ' +
            classes[top_preds[i]])