From 05afc729f2067e9a90be7d61c3e33ea1a5fbb010 Mon Sep 17 00:00:00 2001
From: sushilgoyal1970 <56231544+sushilgoyal1970@users.noreply.github.com>
Date: Sun, 6 Oct 2019 23:13:27 +0530
Subject: [PATCH] creating dataset

---
 dataset_creator.py | 131 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 131 insertions(+)
 create mode 100644 dataset_creator.py

diff --git a/dataset_creator.py b/dataset_creator.py
new file mode 100644
index 0000000..dea164b
--- /dev/null
+++ b/dataset_creator.py
@@ -0,0 +1,131 @@
+import cv2,os,pickle
+import face_recognition as fr
+
+#creating blank lists
+known_face_encodings_list=[]
+known_names=[]
+ids=[]
+font=cv2.FONT_HERSHEY_SIMPLEX
+
+#creating image's directory
+try:
+    cwd=os.getcwd()
+    print(cwd)
+    os.mkdir(cwd+"/dataset_images")
+except:
+    print()
+
+#funtions for Capturing image
+def image_taker(dir_name,student_id):
+    cam=cv2.VideoCapture(0)
+    counter=0
+    flag=0  
+
+    while cam.isOpened():
+        frame=cam.read()[1]
+        
+        #converting BGR frame to RGB frame
+        rgb_frame=frame[:,:,::-1]
+        
+        #getting locations of faces present
+        faces=fr.face_locations(rgb_frame)
+        for (top,right,bottom,left) in faces:
+            cv2.rectangle(frame, (left, top), (right, bottom), (0, 0, 255), 2)
+            
+        cv2.putText(frame, "Press 'C' to capture image", (0 , 35), font, 1.0, (255, 255, 255), 2)
+        cv2.imshow("live",frame)
+            
+            #handler
+        if cv2.waitKey(100) & 0xFF==ord('q'):
+            if flag==0:
+                #remove if image is not created to avoid any issue
+                os.rmdir(dir_name)
+            break
+        if cv2.waitKey(100) & 0xFF==ord('c'):
+            
+            #saving imaegs
+            cv2.imwrite(dir_name+"/image,"+str(student_id)+","+str(counter)+".jpg",frame)
+            flag=1
+            print("captured")
+            cv2.destroyAllWindows()
+    cam.release()
+    cv2.destroyAllWindows()
+   
+    
+    
+     
+
+#setting up student Id
+choice='y'
+
+#getting last student id and creating next id
+try:
+    with open("ids.txt",'rb') as file_data:
+        labels=pickle.load(file_data)
+    student_id=max(labels)+1
+except FileNotFoundError:
+    student_id=0
+    
+while(choice=='y'):
+    
+    print(student_id)
+
+    student_name=input("Enter student name: ")
+    
+    #defining directory name where images will be stored
+    cwd=os.getcwd()+"/dataset_images/"
+    dir_name=cwd+student_name+","+str(student_id)
+    
+    #using try to avoid error when directory is already present
+    try:
+        os.mkdir(dir_name)
+        print("check")
+        image_taker(dir_name,student_id)
+        print("check")
+    except:
+        print("Student name already exits.")
+    
+    choice=input("Add another:")
+    if choice=='y':
+        student_id=student_id+1
+    
+
+
+#getting face encoding from stored images
+dataset_dir_name=os.getcwd()+"/dataset_images"
+folder_names=os.listdir(dataset_dir_name)
+for i in folder_names:
+    dir_name=dataset_dir_name+"/"+i
+    face_names=os.listdir(dir_name)
+    for face_name in face_names:
+        image_name=dir_name+"/"+face_name
+        print(image_name)
+        
+        
+        #loading images using face_recognition library
+        known_face= fr.load_image_file(image_name)
+        print(known_face)
+        
+        #getting encodings of faces
+        known_face_encoding=fr.face_encodings(known_face)[0]
+        student_name=i.split(",")[0]
+        student_id_in=int(i.split(",")[1])
+        
+        #appending encodings,ids, names into lists
+        known_face_encodings_list.append(known_face_encoding)
+        known_names.append(student_name)
+        ids.append(student_id_in)
+    
+
+print(known_names)
+print(ids)
+
+#storing data in files using pickle
+with open("encodings.txt",'wb') as file_data:
+    pickle.dump(known_face_encodings_list,file_data)
+
+with open("name.txt",'wb') as file_data:
+    pickle.dump(known_names,file_data)
+
+with open("ids.txt",'wb') as file_data:
+    pickle.dump(ids,file_data)