02.Advanced_Hand_Pose_Estimation copy.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Advanced Hand Pose Estimation \n",
    "- Detect Left and Right Hand\n",
    "- Calculate Angles"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Reference : [AI Hand Pose Estimation with MediaPipe | Detect Left and Right Hand + Calculate Angles - Nicholas Renotte]('https://www.youtube.com/watch?v=EgjwKM3KzGU')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```\n",
    "Step 1. Import Libraries\n",
    "Step 2. Load Modules\n",
    "Step 3. Function Definition\n",
    "     3-a. Get Each Hand Label\n",
    "     3-b. Draw Finger Angles\n",
    "     3-c. Image Detection\n",
    "Step 4. Control Web Cam\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 1. Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import mediapipe as mp\n",
    "import cv2\n",
    "import numpy as np \n",
    "import uuid \n",
    "import os"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 2. Load Modules (related to hand recogition)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "mp_drawing = mp.solutions.drawing_utils\n",
    "mp_hands = mp.solutions.hands"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 3. Function Definition"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3-a. Get Each Hand Label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_label(index, hand, results):\n",
    "    output = None\n",
    "    for idx, classification in enumerate(results.multi_handedness):\n",
    "        each_hand = classification.classification[0]\n",
    "        if each_hand.index == index:\n",
    "            label = each_hand.label\n",
    "            score = each_hand.score\n",
    "            text = f'{label} {round(score,2)}'\n",
    "\n",
    "            coords = tuple(np.multiply(\n",
    "                np.array((hand.landmark[mp_hands.HandLandmark.WRIST].x,\n",
    "                          hand.landmark[mp_hands.HandLandmark.WRIST].y)),\n",
    "                [640, 480]).astype(int))\n",
    "            # text position - text code\n",
    "            # print('기존:',coords)\n",
    "            # print('수정:',(int(hand.landmark[mp_hands.HandLandmark.WRIST].x*450),\n",
    "            #              int(hand.landmark[mp_hands.HandLandmark.WRIST].y*200)))\n",
    "            output = text, coords\n",
    "\n",
    "    return output"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3-b. Draw Finger Angles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def draw_finger_angles(image, results, joint_list):\n",
    "    # Loop Through hands\n",
    "    for hand in results.multi_hand_landmarks:\n",
    "        # Loop through joint sets\n",
    "        for joint in joint_list:\n",
    "            a = np.array([hand.landmark[joint[0]].x, hand.landmark[joint[0]].y])\n",
    "            b = np.array([hand.landmark[joint[1]].x, hand.landmark[joint[1]].y])\n",
    "            c = np.array([hand.landmark[joint[2]].x, hand.landmark[joint[2]].y])\n",
    "\n",
    "            radians = np.arctan2(c[1] - b[1], c[0]-b[0]) - np.arctan2(a[1]-b[1], a[0]-b[0])\n",
    "            angle = np.abs(radians*180.0/np.pi)\n",
    "\n",
    "            if angle > 180.0:\n",
    "                angle = 360-angle\n",
    "            \n",
    "            cv2.putText(img=image, \n",
    "                        text=str(round(angle,2)), \n",
    "                        org=tuple(np.multiply(b, [640, 480]).astype(int)), \n",
    "                        fontFace=cv2.FONT_HERSHEY_SIMPLEX,\n",
    "                        fontScale=0.5, \n",
    "                        color=(255,255,255), \n",
    "                        thickness=2, \n",
    "                        lineType=cv2.LINE_AA)\n",
    "            \n",
    "\n",
    "    return image"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3-c. Image Detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Image Detection Function\n",
    "def detections(frame):\n",
    "    # BGR to RGB\n",
    "    image = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)\n",
    "\n",
    "    image = cv2.flip(image, 1)\n",
    "\n",
    "    # Set flag\n",
    "    image.flags.writeable = False\n",
    "\n",
    "    # Detection\n",
    "    results = hands.process(image)\n",
    "\n",
    "    # Set flag to true\n",
    "    image.flags.writeable = True\n",
    "\n",
    "    # RGB to BGR\n",
    "    image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)\n",
    "\n",
    "    if results.multi_hand_landmarks:\n",
    "        for num, hand in enumerate(results.multi_hand_landmarks):\n",
    "            mp_drawing.draw_landmarks(image = image, \n",
    "                                      landmark_list = hand, \n",
    "                                      connections = mp_hands.HAND_CONNECTIONS,\n",
    "                                      landmark_drawing_spec = mp_drawing.DrawingSpec(color=(250,250,250), thickness=2, circle_radius=4),\n",
    "                                      connection_drawing_spec = mp_drawing.DrawingSpec(color=(250,250,250), thickness=2, circle_radius=2),\n",
    "                                      )\n",
    "            if get_label(num , hand, results):\n",
    "                text, coord = get_label(num, hand, results)\n",
    "                cv2.putText(img=image, \n",
    "                            text=text, \n",
    "                            org=coord, \n",
    "                            fontFace=cv2.FONT_HERSHEY_SIMPLEX,\n",
    "                            fontScale=1, \n",
    "                            color=(255,255,255), \n",
    "                            thickness=2, \n",
    "                            lineType=cv2.LINE_AA)\n",
    "\n",
    "        # Draw angles to image from joint list\n",
    "        joint_list = [[8,7,6], [12,11,10], [4,3,2], [1,0,5]]\n",
    "        draw_finger_angles(image, results, joint_list)\n",
    "            \n",
    "    return image, results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 4. Control Web Cam "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Control Main Cam\n",
    "cap = cv2.VideoCapture(0)\n",
    "\n",
    "with mp_hands.Hands(min_detection_confidence=0.8, min_tracking_confidence=0.5) as hands:\n",
    "    while cap.isOpened():\n",
    "        ret, frame = cap.read() # ret = True or False, frame : one image from video\n",
    "\n",
    "        if ret == True: # If image exists\n",
    "            image, results = detections(frame) # Draw Landmarks on Hands\n",
    "        \n",
    "        cv2.imshow(\"AI Hand Pose Estimation\", image)\n",
    "        if cv2.waitKey(1) & 0xFF == ord('q'):\n",
    "            break\n",
    "\n",
    "cap.release()\n",
    "cv2.destroyAllWindows()\n",
    "for i in range(10): # This loop was added to normally terminate the python cv2 file\n",
    "    cv2.waitKey(1)  # https://stackoverflow.com/questions/6116564/destroywindow-does-not-close-window-on-mac-using-python-and-opencv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.7.7 64-bit ('dataScience': conda)",
   "name": "python377jvsc74a57bd088212c6b79be9b544bafe6735d74fe8c206b1cc2a4f62e98b323686df1ed5be2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  },
  "orig_nbformat": 2
 },
 "nbformat": 4,
 "nbformat_minor": 2
}