Update app.py

app.py

@@ -1,115 +1,343 @@
-from ultralytics import YOLO
-import random
-import streamlit as st
-from PIL import Image
-import os
-import gdown
-
-def extract_user_input(image=None):
-
-
-    id = "1e2pHejcp96q-NT9VmIDEY_hCA-tiFNef"
-    output = "trained_model.pt"
-    gdown.download(id = id, output = output, quiet=False)
-
-    model = YOLO(output)
-
-    result = model.predict(image)
-
-    names = model.names
-    det_cls = []
-    for r in result:
-        for c in r.boxes.cls:
-            det_cls.append(names[int(c)])
-    
-    return det_cls[0]
-
-
-def ai_choice():
-    choice = random.randint(0,1000000000)
-
-    if choice <=100000:
-        choice = "Rock"
-    elif choice >101 and choice <=1000000:
-        choice = "Paper"
-    else:
-        choice = "Scissors"
-    return choice
-
-def winner(user_input,ai_input):
-
-    if user_input == "Rock":
-        if ai_input == "Scissors":
-            return "Player 1"
-        elif ai_input == "Paper":
-            return "AI"
-        else:
-            return "Draw!!!"
-    if user_input == "Paper":
-        if ai_input == "Rock":
-            return "Player 1"
-        elif ai_input == "Scissors":
-            return "AI"
-        else:
-            return "Draw!!!"
-    if user_input == "Scissors":
-        if ai_input == "Rock":
-            return "AI"
-        elif ai_input == "Paper":
-            return "Player 1"
-        else:
-            return "Draw!!!"
-        
+import cv2
+import mediapipe as mp
+import pyautogui
+import math
+import time
+
+
+class HandMouse:
+    def __init__(self):
+        self.hands = mp.solutions.hands.Hands(min_detection_confidence=0.7, min_tracking_confidence=0.7)
+        self.mp_draw = mp.solutions.drawing_utils
+        self.prev_x, self.prev_y = 0, 0
+        self.smooth_factor = 5
+        self.last_click_time = 0
+        self.last_scroll_time = 0
+
+
+    def find_hands(self, img):
+        img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+        results = self.hands.process(img_rgb)
+        return results
+
+
+    def get_landmarks(self, img, results):
+        lm_list = {}
+        if results.multi_hand_landmarks:
+            for hand_landmarks in results.multi_hand_landmarks:
+                for id, lm in enumerate(hand_landmarks.landmark):
+                    h, w, _ = img.shape
+                    cx, cy = int(lm.x * w), int(lm.y * h)
+                    lm_list[id] = (cx, cy)
+                self.mp_draw.draw_landmarks(img, hand_landmarks, mp.solutions.hands.HAND_CONNECTIONS)
+        return lm_list
+
+
+    def move_mouse(self, img, lm_list):
+        if 8 in lm_list:  # Index finger tip
+            index_tip = lm_list[8]
+            screen_width, screen_height = pyautogui.size()
+            img_width, img_height = img.shape[1], img.shape[0]
+            x, y = index_tip
+
+
+            # Map coordinates to screen size
+            screen_x = int((x / img_width) * screen_width)
+            screen_y = int((y / img_height) * screen_height)
+
+
+            # Smooth mouse movement
+            smoothed_x = self.prev_x + (screen_x - self.prev_x) / self.smooth_factor
+            smoothed_y = self.prev_y + (screen_y - self.prev_y) / self.smooth_factor
+
+
+            pyautogui.moveTo(smoothed_x, smoothed_y)
+            self.prev_x, self.prev_y = smoothed_x, smoothed_y
+
+
+    def left_click(self, lm_list):
+        if 4 in lm_list and 8 in lm_list:  # Thumb and index finger
+            thumb_tip = lm_list[4]
+            index_tip = lm_list[8]
+            distance = math.hypot(thumb_tip[0] - index_tip[0], thumb_tip[1] - index_tip[1])
+
+
+            if distance < 30:  # Click threshold
+                current_time = time.time()
+                if current_time - self.last_click_time > 0.5:  # Prevent multiple clicks
+                    pyautogui.click()
+                    self.last_click_time = current_time
+
+
+    def right_click(self, lm_list):
+        if 8 in lm_list and 12 in lm_list:  # Index and middle finger
+            index_tip = lm_list[8]
+            middle_tip = lm_list[12]
+            distance = math.hypot(index_tip[0] - middle_tip[0], index_tip[1] - middle_tip[1])
+
+
+            if distance < 50:  # Right-click threshold
+                pyautogui.click(button='right')
+                time.sleep(0.5)
+
+
+    def scroll(self, lm_list):
+        if 12 in lm_list and 0 in lm_list:  # Middle finger
+            middle_tip = lm_list[12]
+            palm_base = lm_list[0]
+
+
+            # Calculate vertical distance
+            scroll_distance = middle_tip[1] - palm_base[1]
+            current_time = time.time()
+
+
+            if abs(scroll_distance) > 30 and current_time - self.last_scroll_time > 0.1:
+                if scroll_distance > 0:
+                    pyautogui.scroll(-5)  # Scroll down
+                else:
+                    pyautogui.scroll(5)  # Scroll up
+                self.last_scroll_time = current_time
+
+
+    def copy_content(self, lm_list):
+        if 4 in lm_list and 8 in lm_list and 12 in lm_list:  # Thumb, index, and middle finger
+            thumb_tip = lm_list[4]
+            index_tip = lm_list[8]
+            middle_tip = lm_list[12]
+
+
+            distance_thumb_index = math.hypot(thumb_tip[0] - index_tip[0], thumb_tip[1] - index_tip[1])
+            distance_index_middle = math.hypot(index_tip[0] - middle_tip[0], index_tip[1] - middle_tip[1])
+
+
+            if distance_thumb_index < 30 and distance_index_middle < 30:
+                pyautogui.hotkey('ctrl', 'c')  # Copy content
+                time.sleep(0.5)
+
+
+    def detect_gestures(self, img, lm_list):
+        self.move_mouse(img, lm_list)
+        self.left_click(lm_list)
+        self.right_click(lm_list)
+        self.scroll(lm_list)
+        self.copy_content(lm_list)
+
 
 def main():
-    col1, col2 = st.columns(2)
-    with st.container():
-        st.header("Welcome to the Computer Vison powered Rock-Paper-Scissors Game!")
-        st.caption("developed by Indranil Bhattacharyya")
-
-        player_name = st.text_input("Please enter your name, Player!: ")
-    
-    image = st.camera_input("Please upload your image: ")
-
-    if image:
-        st.image(image)
-
-        im = Image.open(image)
-        rgb_im = im.convert("RGB")
-
-        # exporting the image
-        rgb_im.save("user_input.jpg")
-                
-
-        user_input = extract_user_input("user_input.jpg")
-        user_input_front_end = player_name + " has selected: " + user_input
-
-        st.code(user_input_front_end)
-        ai_choice_var = ai_choice()
-        ai_choice_front_end = "AI has selected: " + ai_choice_var
-        st.code(ai_choice_front_end)
-        if len(user_input)>0:
-            winner_name = winner(user_input,ai_choice_var)
-            if len(winner_name)!=0:
-                with st.container():
-                    if winner_name == "Player 1":
-                       
-                        winner_front_end = player_name + " Wins!!!"
-                        
-                        st.success(winner_front_end)
-                    else:
-                        if winner_name == "AI":
-                            winner_front_end = winner_name + " Wins!!!"
-                        else:
-                            winner_front_end = "It's a " + winner_name
-            
-                        st.error(winner_front_end)
-        else:
-            winner_name = ''
-
-    
+    cap = cv2.VideoCapture(0)
+    hand_mouse = HandMouse()
+
+
+    while True:
+        success, img = cap.read()
+        if not success:
+            break
+
+
+        results = hand_mouse.find_hands(img)
+        lm_list = hand_mouse.get_landmarks(img, results)
+
+
+        if lm_list:
+            hand_mouse.detect_gestures(img, lm_list)
+
+
+        # Show video feed
+        cv2.imshow("Hand Tracking Mouse", img)
+        if cv2.waitKey(1) & 0xFF == ord('q'):
+            break
+
+
+    cap.release()
+    cv2.destroyAllWindows()
+
 
 if __name__ == "__main__":
-    main()       
-        
+    main()
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+import cv2
+
+
+          ######## destop mouse
+import mediapipe as mp
+import pyautogui
+import math
+import time
+
+
+
+
+class HandMouse:
+    def __init__(self):
+        # Initialize Mediapipe Hand Detection
+        self.hands = mp.solutions.hands.Hands(min_detection_confidence=0.7, min_tracking_confidence=0.7)
+        self.mp_draw = mp.solutions.drawing_utils
+
+
+        # Smoothing mouse movement
+        self.prev_x, self.prev_y = 0, 0
+        self.smooth_factor = 5
+
+
+        # Timers to prevent multiple actions
+        self.last_click_time = 0
+        self.last_right_click_time = 0
+        self.last_scroll_time = 0
+
+
+    def find_hands(self, img):
+        """Detect hands and return results."""
+        img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+        results = self.hands.process(img_rgb)
+        return results
+
+
+    def get_landmarks(self, img, results):
+        """Extract hand landmarks."""
+        lm_list = {}
+        if results.multi_hand_landmarks:
+            for hand_landmarks in results.multi_hand_landmarks:
+                for id, lm in enumerate(hand_landmarks.landmark):
+                    h, w, _ = img.shape
+                    cx, cy = int(lm.x * w), int(lm.y * h)
+                    lm_list[id] = (cx, cy)
+                self.mp_draw.draw_landmarks(img, hand_landmarks, mp.solutions.hands.HAND_CONNECTIONS)
+        return lm_list
+
+
+    def move_mouse(self, img, lm_list):
+        """Move the mouse using the index finger."""
+        if 8 in lm_list:  # Index finger tip
+            index_tip = lm_list[8]
+            screen_width, screen_height = pyautogui.size()
+            img_width, img_height = img.shape[1], img.shape[0]
+            x, y = index_tip
+
+
+            # Map to screen size
+            screen_x = int((x / img_width) * screen_width)
+            screen_y = int((y / img_height) * screen_height)
+
+
+            # Smooth the movement
+            smoothed_x = self.prev_x + (screen_x - self.prev_x) / self.smooth_factor
+            smoothed_y = self.prev_y + (screen_y - self.prev_y) / self.smooth_factor
+
+
+            pyautogui.moveTo(smoothed_x, smoothed_y)
+            self.prev_x, self.prev_y = smoothed_x, smoothed_y
+
+
+    def left_click(self, lm_list):
+        """Perform a left click when thumb and index finger touch."""
+        if 4 in lm_list and 8 in lm_list:  # Thumb and index finger
+            thumb_tip = lm_list[4]
+            index_tip = lm_list[8]
+            distance = math.hypot(thumb_tip[0] - index_tip[0], thumb_tip[1] - index_tip[1])
+
+
+            if distance < 30:  # Click threshold
+                current_time = time.time()
+                if current_time - self.last_click_time > 0.5:  # Prevent multiple clicks
+                    pyautogui.click()
+                    self.last_click_time = current_time
+
+
+    def right_click(self, lm_list):
+        """Perform a right click when the thumb finger is raised."""
+        if 4 in lm_list and 0 in lm_list:  # Thumb and palm base
+            thumb_tip = lm_list[4]
+            palm_base = lm_list[0]
+            distance = math.hypot(thumb_tip[0] - palm_base[0], thumb_tip[1] - palm_base[1])
+
+
+            if distance > 50:  # Right-click threshold
+                current_time = time.time()
+                if current_time - self.last_right_click_time > 0.5:  # Prevent multiple clicks
+                    pyautogui.click(button='right')
+                    self.last_right_click_time = current_time
+
+
+    def scroll(self, lm_list):
+        """Scroll up or down using the middle finger."""
+        if 12 in lm_list and 0 in lm_list:  # Middle finger and palm base
+            middle_tip = lm_list[12]
+            palm_base = lm_list[0]
+
+
+            # Calculate vertical distance
+            scroll_distance = middle_tip[1] - palm_base[1]
+            current_time = time.time()
+
+
+            if abs(scroll_distance) > 30 and current_time - self.last_scroll_time > 0.1:
+                if scroll_distance > 0:
+                    pyautogui.scroll(-5)  # Scroll down
+                else:
+                    pyautogui.scroll(5)  # Scroll up
+                self.last_scroll_time = current_time
+
+
+    def detect_gestures(self, img, lm_list):
+        """Detect gestures and perform actions."""
+        self.move_mouse(img, lm_list)
+        self.left_click(lm_list)
+        self.right_click(lm_list)
+        self.scroll(lm_list)
+
+
+
+
+def main():
+    cap = cv2.VideoCapture(0)
+    hand_mouse = HandMouse()
+
+
+    while True:
+        success, img = cap.read()
+        if not success:
+            break
+
+
+        results = hand_mouse.find_hands(img)
+        lm_list = hand_mouse.get_landmarks(img, results)
+
+
+        if lm_list:
+            hand_mouse.detect_gestures(img, lm_list)
+
+
+        # Show video feed
+        cv2.imshow("Hand Tracking Mouse", img)
+        if cv2.waitKey(1) & 0xFF == ord('q'):
+            break
+
+
+    cap.release()
+    cv2.destroyAllWindows()
+
+
+
+
+if __name__ == "__main__":
+    main()
+
+