Awesome

LEG(Linearly Estimated Gradient)

v1.0

Getting Started

Prerequisites

Please make sure that the following packages are installed.

Documentation

methods/LEGv0.py: Implementation of LEG explainer.
ImageNetExp.py: Create the 500 LEG and LEG-TV explanations on ImageNet.
Sanity/: folder including implementations of cascading randomizations on LeNet-5 in MNIST and VGG-19 in ImageNet dataset.
Plots/ : Implementations of plots in papers including sensitivity analysis.
table/keysizetable.py:Implementation of computing key size for each saliency method.

Usage

The LEG_explainer function is called with the following basic inputs:

 LEG_explainer(inputs, model, predict_func, penalty, noise_lvl, lambda_arr, num_sample):

The function returns lists for all inputs. Each list contains the saliency map, original image, prediction of original image and corresponding lambda level for saliency map in turn.

We also provide a customized function for visualization:

generateHeatmap(image, heatmap, name, style, show_option, direction):

You can choose the "heatmap_only", "gray" or "overlay" style for the heatmap and decide whether display original saliency or its absolute value by the direction option.

Following is a toy example:

##Import the required packages
from methods.LEGv0 import * 

if __name__ == "__main__":
    print("We are excuting LEG program", __name__)
    # read the image
    img = image.load_img('Image/trafficlight.jpg', target_size=(224,224))
    img = image.img_to_array(img).astype(int)
    image_input = np.expand_dims(img.copy(), axis = 0)
    image_input = preprocess_input(image_input)
    print("Image has been read successfully")

    # read the model
    VGG19_MODEL = VGG19(include_top = True)
    print("VGG19 has been imported successfully")
    # make the prediction of the image by the vgg19
    preds = VGG19_MODEL.predict(image_input)
    for pred_class in decode_predictions(preds)[0]:
        print(pred_class)
    chosen_class = np.argmax(preds)
    print("The Classfication Category is ", chosen_class)
    begin_time = time()

    LEG = LEG_explainer(np.expand_dims(img.copy(), axis = 0), VGG19_MODEL, predict_vgg19, num_sample = 10000, penalty=None)
    LEGTV = LEG_explainer(np.expand_dims(img.copy(), axis = 0), VGG19_MODEL, predict_vgg19, num_sample = 10000, penalty='TV', lambda_arr = [0.1, 0.3])
    end_time = time()
    plt.imshow(LEG[0][0], cmap='hot', interpolation="nearest")
    plt.show() #change the backend of matplotlib if it can not be displayed 
    plt.imshow(LEGTV[0][0], cmap='hot', interpolation="nearest")
    plt.show()
    
    print("The time spent on LEG explanation is ",round((end_time - begin_time)/60,2), "mins")

For the choice of parameters, we suggest using [0.02] as the noise level, and using [0.1, 0.3] as lambda values. Plesae specify the path for inputs and outputs before execution for reproducibility.