improving code

README.md

@@ -20,7 +20,7 @@ In practice, a **higher mAP** value indicates a **better performance** of your n
 ## Explanation
 The performance of your neural net will be judged using the mAP criterium defined in the [PASCAL VOC 2012 competition](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/). We simply adapted the [official Matlab code](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/#devkit) into Python (in our tests they both give the same results).
 
-First (**1.**), we calculate the Average Precision (AP), for each of the classes present in the ground-truth. Then (**2.**), we calculate the mean of all the AP's, resulting in a mAP value.
+First (**1.**), we calculate the Average Precision (AP), for each of the classes present in the ground-truth. Then (**2.**), we calculate the mean of all the AP's, resulting in an mAP value.
 
 #### 1. Calculate AP for each Class
 
@@ -61,7 +61,7 @@ Step by step:
 
 Optional (if you want to see the **animation**):
 
-  6. Insert the images in the folder **images/**
+  6. Insert the images into the folder **images/**
 
 
 #### Create the ground-truth files

extra/README.md

@@ -1,6 +1,13 @@
 # Extra
 
-### Convert YOLO to PASCAL VOC format:
+## Ground-Truth:
+- ### convert YOLO to PASCAL VOC format:
 
-1) Add class list to the file `class_list.txt`
-2) Run the python script and specify the path as an argument, e.g. `python convert_yolo_format_to_voc.py ../ground-truth`
+    1) Add class list to the file `class_list.txt`
+    2) Insert ground-truth files into **ground-truth/**
+    3) Insert images into **images/**
+    4) Run the python script: `python convert_yolo_format_to_voc.py`
+
+## Find the files that contain a specific class of objects
+
+1) Run the python script and specify the **class** and **path** as arguments, e.g. `python find_class.py chair ../ground-truth`

extra/convert_yolo_format_to_voc.py

@@ -13,7 +13,7 @@
 #print(obj_list[0])
 
 # change directory to the one with the files to be changed
-path_to_folder = sys.argv[1]
+path_to_folder = '../ground_truth'
 #print(path_to_folder)
 os.chdir(path_to_folder)
 

extra/find_class.py

@@ -0,0 +1,25 @@
+import sys
+import os
+import glob
+
+if len(sys.argv) != 3:
+  print("Error: wrong format.\nUsage: python find_class.py [class_name] [path]")
+  sys.exit(0)
+
+searching_class_name = sys.argv[1]
+path = sys.argv[2]
+
+os.chdir(path)
+
+# iterate through the text files
+for txt_file in glob.glob('*.txt'):
+  # open txt file lines to a list
+  with open(txt_file) as f:
+    content = f.readlines()
+  # remove whitespace characters like `\n` at the end of each line
+  content = [x.strip() for x in content]
+  # go through each line of eache file
+  for line in content:
+    class_name = line.split()[0]
+    if class_name == searching_class_name:
+      print(txt_file)

main.py

@@ -3,6 +3,7 @@
 import os
 import shutil
 import operator
+import sys
 import argparse
 
 parser = argparse.ArgumentParser()
@@ -17,15 +18,34 @@
 group.add_argument('-fa', '--fast-animation', help="animation shown fast.", action="store_true")
 args = parser.parse_args()
 
+# if there are no classes to ignore then replace None by empty list
 if args.ignore is None:
   args.ignore = []
 
+# if there are no images then no animation can be shown
+img_path = 'images'
+if os.path.exists(img_path): 
+  for dirpath, dirnames, files in os.walk(img_path):
+    if not files:
+      # no image files found
+      args.no_animation = True
+else:
+  args.no_animation = True
+
+# try to import OpenCV if the user didn't choose the option --no-animation
 if not args.no_animation:
-  import cv2
+  try:
+    import cv2
+  except ImportError:
+    args.no_animation = True
 
+# try to import Matplotlib if the user didn't choose the option --no-plot
 if not args.no_plot:
-  import matplotlib.pyplot as plt
-  import numpy as np
+  try:
+    import matplotlib.pyplot as plt
+    import numpy as np
+  except ImportError:
+    args.no_plot = True
 
 MINOVERLAP = 0.5 # value defined in the PASCAL VOC2012 challenge
 
@@ -94,7 +114,7 @@ def file_lines_to_list(path):
 """
  Ground-Truth
    Load each of the ground-truth files into a temporary ".json" file.
-   Create a list with all the class names present in the ground-truth (unique_classes).
+   Create a list of all the class names present in the ground-truth (unique_classes).
 """
 # get a list with the ground-truth files
 ground_truth_files_list = glob.glob('ground-truth/*.txt')
@@ -104,10 +124,16 @@ def file_lines_to_list(path):
 counter_per_class = {}
 for txt_file in ground_truth_files_list:
   #print(txt_file)
-  lines = file_lines_to_list(txt_file)
+  file_id = txt_file.split(".txt",1)[0]
+  file_id = file_id.split("/",1)[1]
+  # check if there is a correspondent predicted objects file
+  if not os.path.exists('predicted/' + file_id + ".txt"):
+    print("Error. File not found: predicted/" +  file_id + ".txt")
+    sys.exit(0)
+  lines_list = file_lines_to_list(txt_file)
   # create ground-truth dictionary
   bounding_boxes = []
-  for line in lines:
+  for line in lines_list:
     class_name, left, top, right, bottom = line.split()
     # check if class is in the ignore list, if yes skip
     if class_name in args.ignore:
@@ -122,8 +148,6 @@ def file_lines_to_list(path):
       counter_per_class[class_name] = 1
       unique_classes.add(class_name)
   # dump bounding_boxes into a ".json" file
-  file_id = txt_file.split(".txt",1)[0]
-  file_id = file_id.split("/",1)[1]
   with open(tmp_files_path + "/" + file_id + "_ground_truth.json", 'w') as outfile:
     json.dump(bounding_boxes, outfile)
 
@@ -207,15 +231,17 @@ def file_lines_to_list(path):
   fp = [0] * nd
   for idx, prediction in enumerate(predictions_data):
     file_id = prediction["file_id"]
-    # find ground truth image
-    ground_truth_img = glob.glob1("images", file_id + "*")
-    #tifCounter = len(glob.glob1(myPath,"*.tif"))
-    if len(ground_truth_img) == 0:
-      print("Error: unrecognized image " + file_id)
-      sys.exit(0)
-    elif len(ground_truth_img) > 1:
-      print("Error: multiple image " + file_id)
-      sys.exit(0)
+    # if show animation
+    if not args.no_animation:
+      # find ground truth image
+      ground_truth_img = glob.glob1(img_path, file_id + ".*")
+      #tifCounter = len(glob.glob1(myPath,"*.tif"))
+      if len(ground_truth_img) == 0:
+        print("Error. Image not found with id: " + file_id)
+        sys.exit(0)
+      elif len(ground_truth_img) > 1:
+        print("Error. Multiple image with id: " + file_id)
+        sys.exit(0)
     # assign prediction to ground truth object if any
     #   open ground-truth with that file_id
     gt_file = tmp_files_path + "/" + file_id + "_ground_truth.json"