added training data generator files

.gitignore

@@ -2,4 +2,10 @@
 venv/*
 output/*
 .envrc
-training_data/*
+training_data/data*
+training_data/info*
+training_data/training_data_*/
+training_data/*.vec
+training_data/backgrounds.txt
+training_data/negatives
+training_data/opencv

Main.py

@@ -82,12 +82,7 @@ if args.file:
     cap = cv2.VideoCapture(args.file)
 else:
     cap = cv2.VideoCapture(0, cv2.IMREAD_GRAYSCALE) # instead of grayscale you can also use -1, 0, or 1.
-faceCascade = cv2.CascadeClassifier(r"./cascades/lbpcascade_frontalface.xml") # CHECK THIS FIRST TROUBLE SHOOTING
+faceCascade = cv2.CascadeClassifier(r"./cascades/cascade_10.xml") # CHECK THIS FIRST TROUBLE SHOOTING
-faceCascade_default = cv2.CascadeClassifier(r"./cascades/haarcascade_frontalface_default.xml")
-faceCascade_alt = cv2.CascadeClassifier(r"./cascades/haarcascade_frontalface_alt.xml")
-faceCascade_alt2 = cv2.CascadeClassifier(r"./cascades/haarcascade_frontalface_alt2.xml")
-faceCascade_alttree = cv2.CascadeClassifier(r"./cascades/haarcascade_frontalface_alt_tree.xml")
-profileFaceCascade = cv2.CascadeClassifier(r"./cascades/haarcascade_profileface.xml")
 
 datestamp = "{:%Y_%m_%d %H_%M_%S}".format(datetime.datetime.now())
 output_dir = r"./output/" + datestamp + r"/"
@@ -112,51 +107,12 @@ while(True):
     # Detect faces in the image
     faces = faceCascade.detectMultiScale(
     gray,
-    scaleFactor=1.1,
+    scaleFactor=1.2,
-    minNeighbors=5,
+    minNeighbors=2,
-    minSize=(30, 30)
+    # minSize=(70, 90)
+    minSize=(200, 200)
     )
 
-    if len(faces) == 0:
-        faces = faceCascade_default.detectMultiScale(
-            gray,
-            scaleFactor=1.1,
-            minNeighbors=5,
-            minSize=(30,30)
-        )
-
-    if len(faces) == 0:
-        faces = profileFaceCascade.detectMultiScale(
-            gray,
-            scaleFactor=1.1,
-            minNeighbors=5,
-            minSize=(30,30)
-        )
-
-    if len(faces) == 0:
-        faces = faceCascade_alt.detectMultiScale(
-            gray,
-            scaleFactor=1.1,
-            minNeighbors=5,
-            minSize=(30,30)
-        )
-
-    if len(faces) == 0:
-        faces = faceCascade_alt2.detectMultiScale(
-            gray,
-            scaleFactor=1.1,
-            minNeighbors=5,
-            minSize=(30,30)
-        )
-
-    if len(faces) == 0:
-        faces = faceCascade_alttree.detectMultiScale(
-            gray,
-            scaleFactor=1.1,
-            minNeighbors=5,
-            minSize=(30,30)
-        )
-
     # Draw a rectangle around the faces
     for (x, y, w, h) in faces:
         if args.training_data: 

quickstart.txt

@@ -17,4 +17,9 @@ Now you can run the program. It is recommended to run the program with -d and -o
 
 
 Training Data:
-https://www.kaggle.com/datasets/utkarshsaxenadn/landscape-recognition-image-dataset-12k-images
+https://www.kaggle.com/datasets/utkarshsaxenadn/landscape-recognition-image-dataset-12k-images
+
+
+create positives from the negatives: \opencv\build\x64\vc15\bin\opencv_createsamples.exe -img .\positives\face_1.png -bg .\bg.txt -info info/info.lst -pngoutput info -maxxangle 0.8 -maxyangle 0.8 -maxzangle 0.8 -num 1950
+Create vec files from positives: .\opencv\build\x64\vc15\bin\opencv_createsamples.exe -info .\info\info.lst -num 1950 -w 80 -h 80 -vec positives-80.vec
+(I created a 20, 40, and 80) we have 1650 positives

training_data/create_ground_truth.py

@@ -0,0 +1,93 @@
+import cv2
+import sys
+ 
+(major_ver, minor_ver, subminor_ver) = (cv2.__version__).split('.')
+ 
+if __name__ == '__main__' :
+ 
+    # Set up tracker.
+    # Instead of MIL, you can also use
+ 
+    tracker_types = ['BOOSTING', 'MIL','KCF', 'TLD', 'MEDIANFLOW', 'GOTURN', 'MOSSE', 'CSRT']
+    tracker_type = tracker_types[1]
+ 
+    if int(minor_ver) < 3:
+        tracker = cv2.Tracker_create(tracker_type)
+    else:
+        if tracker_type == 'BOOSTING':
+            tracker = cv2.TrackerBoosting_create()
+        if tracker_type == 'MIL':
+            tracker = cv2.TrackerMIL_create()
+        if tracker_type == 'KCF':
+            tracker = cv2.TrackerKCF_create()
+        if tracker_type == 'TLD':
+            tracker = cv2.TrackerTLD_create()
+        if tracker_type == 'MEDIANFLOW':
+            tracker = cv2.TrackerMedianFlow_create()
+        if tracker_type == 'GOTURN':
+            tracker = cv2.TrackerGOTURN_create()
+        if tracker_type == 'MOSSE':
+            tracker = cv2.TrackerMOSSE_create()
+        if tracker_type == "CSRT":
+            tracker = cv2.TrackerCSRT_create()
+ 
+    # Read video
+    video = cv2.VideoCapture("./TestVideo.mp4")
+ 
+    # Exit if video not opened.
+    if not video.isOpened():
+        print("Could not open video")
+        sys.exit()
+ 
+    # Read first frame.
+    ok, frame = video.read()
+    if not ok:
+        print('Cannot read video file')
+        sys.exit()
+     
+    # Define an initial bounding box
+    bbox = (287, 23, 86, 320)
+ 
+    # Uncomment the line below to select a different bounding box
+    bbox = cv2.selectROI(frame, False)
+ 
+    # Initialize tracker with first frame and bounding box
+    ok = tracker.init(frame, bbox)
+ 
+    while True:
+        # Read a new frame
+        ok, frame = video.read()
+        if not ok:
+            break
+         
+        # Start timer
+        timer = cv2.getTickCount()
+ 
+        # Update tracker
+        ok, bbox = tracker.update(frame)
+ 
+        # Calculate Frames per second (FPS)
+        fps = cv2.getTickFrequency() / (cv2.getTickCount() - timer);
+ 
+        # Draw bounding box
+        if ok:
+            # Tracking success
+            p1 = (int(bbox[0]), int(bbox[1]))
+            p2 = (int(bbox[0] + bbox[2]), int(bbox[1] + bbox[3]))
+            cv2.rectangle(frame, p1, p2, (255,0,0), 2, 1)
+        else :
+            # Tracking failure
+            cv2.putText(frame, "Tracking failure detected", (100,80), cv2.FONT_HERSHEY_SIMPLEX, 0.75,(0,0,255),2)
+ 
+        # Display tracker type on frame
+        cv2.putText(frame, tracker_type + " Tracker", (100,20), cv2.FONT_HERSHEY_SIMPLEX, 0.75, (50,170,50),2);
+     
+        # Display FPS on frame
+        cv2.putText(frame, "FPS : " + str(int(fps)), (100,50), cv2.FONT_HERSHEY_SIMPLEX, 0.75, (50,170,50), 2);
+ 
+        # Display result
+        cv2.imshow("Tracking", frame)
+ 
+        # Exit if ESC pressed
+        k = cv2.waitKey(1) & 0xff
+        if k == 27 : break

training_data/training_data_setup.py

@@ -0,0 +1,120 @@
+from PIL import Image
+import os
+import subprocess
+import shutil
+
+backgrounds_file_path = "backgrounds.txt"
+info_base_path = r"./info"
+negatives_path = r"./negatives"
+positives_path = r"./positives"
+training_data_base = r"./training_data_"
+
+opencv_path = r".\opencv\build\x64\vc15\bin\opencv_createsamples.exe"
+
+set_sizes = [1, 2, 5, 10]
+
+max_xangle = 0.5
+max_yangle = 0.5
+max_zangle = 0.5
+
+w, h = 25, 18
+
+class InfoEntry:
+    info_lst_line: str 
+    image_path: str
+
+    def __init__(self, info_line, file_path):
+        self.info_lst_line = info_line 
+        self.image_path = file_path
+
+    def __str__(self):
+        return f"Image Entry: {self.info_lst_line}, {self.image_path}"
+
+
+max_x = 750
+max_y = 800
+
+# remove too small images
+for image in os.listdir("./negatives"):
+    im = Image.open(f"./negatives/{image}")
+    width, height = im.size
+    del im
+    if width <= max_x:
+        os.remove(f"./negatives/{image}")
+    elif height <= max_y:
+        os.remove(f"./negatives/{image}")
+
+# remove any existing file and assume old data
+if os.path.exists(backgrounds_file_path):
+    os.remove(backgrounds_file_path)
+
+# regenerate the available negatives list
+count_negatives = len(os.listdir(negatives_path))
+for img in os.listdir(negatives_path):
+    line = f"{negatives_path}/" + img + "\n"
+    with open(backgrounds_file_path, 'a') as f:
+        f.write(line)
+
+info_dirs = []
+
+if len(os.listdir(positives_path)) > max(set_sizes):
+    print("Your set sizes were larger than the available positive images!")
+    quit(2)
+
+for img in os.listdir(positives_path):
+    i = len(info_dirs)
+    info_dir = f"{info_base_path}{i}"
+
+    com = f"{opencv_path} -img positives/" + str(i) + ".png -bg backgrounds.txt -info " + info_dir + "/info.lst" + \
+    " -pngoutput " + info_dir + " -maxxangle " + str(max_xangle) + " -maxyangle " + str(max_yangle) + " -maxzangle " + str(max_zangle) + \
+    " -num " + str(count_negatives)
+
+    if not os.path.exists(info_dir):
+        subprocess.call(com, shell=True)
+
+    info_dirs.append(info_dir)
+
+for i in set_sizes:
+    if not os.path.exists(training_data_base + str(i)):
+        os.makedirs(training_data_base + str(i))
+
+def join_info_folders(info_dirs: list, output_dir: str):
+    info_dir: str
+    cur_entry_name = 0
+    for info_dir in info_dirs:
+        info_lines = []
+        with open(info_dir + "/info.lst", 'r') as info_file:
+            for line in info_file.readlines():
+                image_path = f"{info_dir}/{line.split(' ')[0]}"
+                info_lines.append(InfoEntry(line.strip(), image_path))
+
+        item: InfoEntry
+        for item in info_lines:
+            shutil.copy(item.image_path, f"{output_dir}/{str(cur_entry_name)}.jpg")
+            with open(f"{output_dir}/info.lst", 'a') as info_file:
+                to_write = []
+                to_write.append(str(cur_entry_name) + ".jpg")
+                to_write = to_write + item.info_lst_line.split(" ")[1:]
+                to_write.append("\n")
+                info_file.write(" ".join(to_write))
+            cur_entry_name += 1
+
+for i in set_sizes:
+    join_info_folders(info_dirs[:i], training_data_base + str(i)) 
+
+commands = []
+
+for i in set_sizes:
+    num_positives = len(os.listdir(training_data_base + str(i)))
+    if os.path.exists(training_data_base + str(i) + ".vec"):
+        os.remove(training_data_base + str(i) + ".vec")
+    com = f"{opencv_path} -info {training_data_base + str(i)}\info.lst -num {num_positives} -w {w} -h {h} -vec {training_data_base + str(i)}.vec"
+    subprocess.call(com, shell=True)
+    commands.append(f".\opencv\\build\\x64\\vc15\\bin\opencv_traincascade.exe -data data_{str(i)} -vec .\\{training_data_base + str(i)}.vec -bg .\\{backgrounds_file_path} -numPos {num_positives} -numNeg {num_positives / 2} -numStages 15 -w {w} -h {h}")
+
+    if not os.path.exists(".\data_" + str(i)):
+        os.makedirs(".\data_" + str(i))
+
+for i in commands:
+    print(f"You are ready to train the models with: \n {i}")
+