CODEWITHSUNDEEP
pythonimage-processingscikit-learnsvm
Ahmed Fawzy Bargal
Ahmed Fawzy Bargal

Reputation: 7

How to get the confidence of svm algorithm in sklearn using python

I want to get the confidence but when I try clf.score(X_test) I get the following error:

TypeError: score() takes at least 3 arguments (2 given)

This is my code:

import cv2, os
from numpy import *
import numpy as np
from PIL import Image
from sklearn.decomposition import PCA
from sklearn.decomposition import RandomizedPCA
from sklearn.naive_bayes import GaussianNB
import matplotlib.pyplot as plt
from sklearn.svm import SVC

cascadeLocation = "haarcascade_frontalface_default.xml"
faceCascade = cv2.CascadeClassifier(cascadeLocation)

def prepare_dataset(directory):
    paths = [os.path.join(directory, filename) for filename in os.listdir(directory)]
    images = []
    labels = []
    row = 140
    col = 140
    for image_path in paths:

        image_pil = Image.open(image_path).convert('L')

        image = np.array(image_pil, 'uint8')
        nbr = int(os.path.split(image_path)[-1].split('.')[1])
        print(nbr)
        faces = faceCascade.detectMultiScale(image)
        for (x,y,w,h) in faces:
            images.append(image[y:y+col,x:x+row])
            labels.append(nbr)
            cv2.imshow("Reading Faces ",image[y:y+col,x:x+row])
            cv2.waitKey(50)
    return images,labels, row, col

directory = 'dataset'
directory2 = 'dataset2'
images, labels, row, col = prepare_dataset(directory)

n_components = 10
cv2.destroyAllWindows()
pca = PCA(n_components=n_components, whiten=True)

param_grid = {'C': [1e3, 5e3, 1e4, 5e4, 1e5],
              'gamma': [0.0001, 0.0005, 0.001, 0.005, 0.01, 0.1], }
clf = SVC(kernel='rbf', class_weight='balanced')

testing_data = []
for i in range(len(images)):
    testing_data.append(images[i].flatten())
pca = pca.fit(testing_data)

transformed = pca.transform(testing_data)
clf.fit(transformed,labels)

image_paths = [os.path.join(directory2, filename) for filename in os.listdir(directory2)]
la=[]
for image_path in image_paths:
    pred_image_pil = Image.open(image_path).convert('L')

    pred_image = np.array(pred_image_pil, 'uint8')

    faces = faceCascade.detectMultiScale(pred_image)
    for (x,y,w,h) in faces:
        temp=np.array(pred_image[y:y+col,x:x+row]).reshape((1, -1))

        X_test = pca.transform(temp)

        id = clf.predict(X_test)
        #a=clf.score(X_test)            
        if(id==1):
            id="john"
        elif(id==2):
            id="brad"
        elif(id==3):
            id="scr"
        elif(id==4):
            id="natalie portman"
        elif(id==5):
            id="jennifer lawrence"
        elif(id==6):
            id="van diesel"
        elif(id==7):
            id="jennifer aniston"
        elif(id==8):
            id="leonardo dicaprio"    
        else :
            id="unknown"
        print(id)               
        cv2.imshow("Recognizing Face", pred_image[y: y + h, x: x + w])
        cv2.waitKey(1000)
cv2.destroyAllWindows()

I also tried clf.score(X_test,id). In this case all confidence takes the value 1.0.

Upvotes: 0

Views: 1866

Answers (1)

Tonechas
Tonechas

Reputation: 13723

From the documentation (emphasis mine):

decision_function(X)
     Predict confidence scores for samples.
     The confidence score for a sample is the signed distance of that sample to the hyperplane.

You have to change a=clf.score(X_test) to a = clf.decision_function(X_test) in your code.

EDIT

I'm guessing you actually wish to "manually" validate your model by checking whether dataset2's images are recognized or not. In that case you need to introduce a number of changes in your code:

  • Put his sentence at the very beginning of the outer looploop:

    nbr2 = int(os.path.split(image_path)[-1].split('.')[1])

    I'm assuming that the file naming conventions for dataset2 are the same than those for dataset1, otherwise you would have to redefine nbr2 accordingly.

  • Don't use id as a variable name since id is a reserved word in Python. You could use idxinstead: 

    idx = clf.predict(X_test)

  • Replace the compound if-elif-else statement by a dictionary (outside the forloop): 

    names = {1: "john",
         2: "brad",
         3: "scr",
         4: "natalie portman",
         5: "jennifer lawrence",
         6: "van diesel", 
         7: "jennifer aniston",
         8: "leonardo dicaprio",
         }

    and change print(id) to print(names.get(idx, "unknown")).

  • And last, replace a=clf.score(X_test) by a = clf.score(X_test, nbr2).

Upvotes: 1

Related Questions