This function will assemble the values of the image directory, and from random images, to perform a clustering on those, and will return the

:param main_label_bottlenecks: Numpy array containing all the bottleneck values of your main label. :param polution_label_bottlenecks: Numpy

We take the assumption that the data set contains less than 50 % of outlier. Given that the classifier, gives the label 0 and 1 for the same

Given a shape you want for your images, will create a mozaic, the size of your screen to visualize multiple image at once. :para

:param main_label_bottlenecks: Numpy array containing all the bottleneck values of your main label. :param polution_label_bottlenecks: Numpy

:param bottlenecks: :param labels: :param suptitle: String to add as plot suptitles :return: Nothing, will just plot a scatter plot

This function will delete our detected images to the desired location. :param image_paths: A list containing the paths to every images detect

Really good if the classes you are analyzing are close to what the network learned. :param image_set: The bottleneck values of the relevant

:param image_set: The bottleneck values of the relevant images. :return: Predictions vector

:param image_set: The bottleneck values of the relevant images. :return: Predictions vector

Fast, but might not be able to map great for nonlinear separation of classes. :param image_set: The bottleneck values of the relevant images

This function will move our detected images to the desired location. :param relocation_dir: The new location for our detected images. :pa

Adds operations that perform JPEG decoding and resizing to the graph.. Args: input_width: Desired width of the image fed into the recognize

Creates a graph from saved GraphDef file and returns a Graph object. Args: model_info: Dictionary containing information about the model ar

Given the name of a model architecture, returns information about it. There are different base image recognition pretrained models that can be

Function that will compute your botlleneck values from your selection of noisy data. :param image_dir: List of image dir location you want t

A simple function that will find the paths of the detected images. :param image_dir: The location of the image directory. :param predicti

A function that returns the paths of images detected, and not yet processed. :param all_paths: Paths of all original, images :param alrea

This function that will see if the pollution directory exist, and try to look a .npy file following the right naming scheme. :param arch

Function that will look if your label as already been transformed to bottleneck given a model, will register it or load, making it easy to ti

Download and extract model tar file. If the pretrained model we're using doesn't already exist, this function downloads it from the TensorFlo

Runs inference on an image to extract the 'bottleneck' summary layer. Args: sess: Current active TensorFlow Session. image_data: Stri

This method will check the values given by the user. :param _: Parser :return: Nothing

This method will check the values given by the user. :param _: Parser :return: Nothing

Creates a graph from saved GraphDef file and returns a Graph object. Args: model_info: Dictionary containing information about the model ar

:param predictions: Vector of labels given by the classifier :param ground_truth: Vector of labels of the data :return: accuracy, precis

Construct and display images that were mislabeled by our classifier :param image_set: Entire set of images :param predictions: Vector of

Construct and display images that were mislabeled by our classifier :param image_set: Entire set of images :param predictions: Vector of