Configuration file loaded

README.md

@@ -11,13 +11,15 @@ test-object-detector
 
 _The `test-object-detector` will download the [UIUC Image Database for Car Detection](https://cogcomp.cs.illinois.edu/Data/Car/) and train a classifier to detect cars in an image. The SVM model files will be stored in `data/models`, so that they can be resused later on._
 
+* All the configurations are in the `data/config/config.cfg` configuration files. You can change it as per your need.
+
 # About the modules
 
 1. `extract-features.py` -- This module is used to extract HOG features of the training images.
 2. `train-classifier.py` -- This module is used to train the classifier.
 3. `nms.py` -- This module performs Non Maxima Suppression.
-4. `test-classifier.py` -- This module is used to test the classifier using a test image
-
+4. `test-classifier.py` -- This module is used to test the classifier using a test image.
+5. `config.py` -- Imports the configuration variables from `config.cfg`.
 
 ## Some of the results
 
@@ -62,3 +64,4 @@ _Detections after NMS_
 4. [Non-Maximum Suppression for Object Detection in Python](http:https://www.pyimagesearch.com/2014/11/17/non-maximum-suppression-object-detection-python/)
 5. [(Faster) Non-Maximum Suppression in Python](http:https://www.pyimagesearch.com/2015/02/16/faster-non-maximum-suppression-python/)
 6. [Texture Matching using Local Binary Patterns (LBP), OpenCV, scikit-learn and Python](http:https://hanzratech.in/2015/05/30/local-binary-patterns.html)
+7. [Detección de objetos Course by Coursera](https://www.coursera.org/course/deteccionobjetos)

data/config/config.cfg

@@ -0,0 +1,17 @@
+[hog]
+min_wdw_sz: [100, 40]
+step_size: [10, 10]
+orientations: 9
+pixels_per_cell: [8, 8]
+cells_per_block: [3, 3]
+visualize: False
+normalize: True
+
+[nms]
+threshold: .3
+
+[paths]
+pos_feat_ph: ../data/features/pos
+neg_feat_ph: ../data/features/neg
+model_path: ../data/models/svm.model
+

object-detector/__init__.py

@@ -2,3 +2,4 @@
 from .train-classifier import *
 from .test-classifier import *
 from .nms import *
+from .config import *

object-detector/config.py

@@ -0,0 +1,21 @@
+'''
+Set the config variable.
+'''
+
+import ConfigParser as cp
+import json
+
+config = cp.RawConfigParser()
+config.read('../data/config/config.cfg')
+
+min_wdw_sz = json.loads(config.get("hog","min_wdw_sz"))
+step_size = json.loads(config.get("hog", "step_size"))
+orientations = config.getint("hog", "orientations")
+pixels_per_cell = json.loads(config.get("hog", "pixels_per_cell"))
+cells_per_block = json.loads(config.get("hog", "cells_per_block"))
+visualize = config.getboolean("hog", "visualize")
+normalize = config.getboolean("hog", "normalize")
+pos_feat_ph = config.get("paths", "pos_feat_ph")
+neg_feat_ph = config.get("paths", "neg_feat_ph")
+model_path = config.get("paths", "model_path")
+threshold = config.getfloat("nms", "threshold")

object-detector/extract-features.py

@@ -7,6 +7,7 @@
 import argparse as ap
 import glob
 import os
+from config import *
 
 if __name__ == "__main__":
  # Argument Parser
@@ -22,9 +23,6 @@
  pos_im_path = args["pospath"]
  neg_im_path = args["negpath"]
 
- pos_feat_ph = "../data/features/pos"
- neg_feat_ph = "../data/features/neg"
-
  des_type = args["descriptor"]
 
  # If feature directories don't exist, create them
@@ -39,7 +37,7 @@
  for im_path in glob.glob(os.path.join(pos_im_path, "*")):
  im = imread(im_path, as_grey=True)
  if des_type == "HOG":
- fd = hog(im, orientations=9, pixels_per_cell=(8,8), cells_per_block=(3,3), visualise=False, normalise=True)
+ fd = hog(im, orientations, pixels_per_cell, cells_per_block, visualize, normalize)
  fd_name = os.path.split(im_path)[1].split(".")[0] + ".feat"
  fd_path = os.path.join(pos_feat_ph, fd_name)
  joblib.dump(fd, fd_path)
@@ -49,7 +47,7 @@
  for im_path in glob.glob(os.path.join(neg_im_path, "*")):
  im = imread(im_path, as_grey=True)
  if des_type == "HOG":
- fd = hog(im, orientations=9, pixels_per_cell=(8,8), cells_per_block=(3,3), visualise=False, normalise=True)
+ fd = hog(im,  orientations, pixels_per_cell, cells_per_block, visualize, normalize)
  fd_name = os.path.split(im_path)[1].split(".")[0] + ".feat"
  fd_path = os.path.join(neg_feat_ph, fd_name)
  joblib.dump(fd, fd_path)

object-detector/test-classifier.py

@@ -6,6 +6,7 @@
 import cv2
 import argparse as ap
 from nms import nms
+from config import *
 
 def sliding_window(image, window_size, step_size):
  '''
@@ -43,10 +44,9 @@ def sliding_window(image, window_size, step_size):
  min_wdw_sz = (100, 40)
  step_size = (10, 10)
  downscale = args['downscale']
- visualize = args['visualize']
+ visualize_det = args['visualize']
 
  # Load the classifier
- model_path = '../data/models/svm.model'
  clf = joblib.load(model_path)
 
  # List to store the detections
@@ -65,9 +65,7 @@ def sliding_window(image, window_size, step_size):
  if im_window.shape[0] != min_wdw_sz[1] or im_window.shape[1] != min_wdw_sz[0]:
  continue
  # Calculate the HOG features
- fd = hog(im_window, orientations=9, pixels_per_cell=(8, 8),
- cells_per_block=(3, 3), visualise=False, normalise=True)
- # Perform the predictions
+ fd = hog(im_window, orientations, pixels_per_cell, cells_per_block, visualize, normalize)
  pred = clf.predict(fd)
  if pred == 1:
  print "Detection:: Location -> ({}, {})".format(x, y)
@@ -78,7 +76,7 @@ def sliding_window(image, window_size, step_size):
  cd.append(detections[-1])
  # If visualize is set to true, display the working
  # of the sliding window
- if visualize:
+ if visualize_det:
  clone = im_scaled.copy()
  for x1, y1, _, _, _ in cd:
  # Draw the detections at this scale
@@ -100,7 +98,7 @@ def sliding_window(image, window_size, step_size):
  cv2.waitKey()
 
  # Perform Non Maxima Suppression
- detections = nms(detections, .3)
+ detections = nms(detections, threshold)
 
  # Display the results after performing NMS
  for (x_tl, y_tl, _, w, h) in detections:

object-detector/train-classifier.py

@@ -6,6 +6,7 @@
 import argparse as ap
 import glob
 import os
+from config import *
 
 if __name__ == "__main__":
  # Parse the command line arguments
@@ -39,7 +40,6 @@
  clf = LinearSVC()
  print "Training a Linear SVM Classifier"
  clf.fit(fds, labels)
- model_path = '../data/models/svm.model'
  # If feature directories don't exist, create them
  if not os.path.isdir(os.path.split(model_path)[0]):
  os.makedirs(os.path.split(model_path)[0])