What is the [concept][1] and how to calculate Bottleneck values ? How these values helps in image classification. Please explain in simple word.


  [1]: https://www.tensorflow.org/tutorials/image_retraining