{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Pedestrian Detection (Default)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2016/03/28 ysmrnbt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2 # use opencv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The HOG detector returns slightly larger rectangles than the real objects.\n",
    "So we slightly shrink the rectangles to get a nicer output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def draw_detections(img, rects, thickness = 1):\n",
    "    for x, y, w, h in rects:\n",
    "        pad_w, pad_h = int(0.15*w), int(0.05*h)\n",
    "        cv2.rectangle(img, (x+pad_w, y+pad_h), (x+w-pad_w, y+h-pad_h), (0, 255, 0), thickness)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Default detector: INRIA Person Dataset is used for training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "HOG(Histogram of Oriented Gradients) and SVM(Support Vector Machine) are used to detect pedestrian in 1 picture.\n",
    "- hitThreshold: Threshold for the distance between features and SVM classifying plane. Usually it is 0 and should be specfied in the detector coefficients (as the last free coefficient). But if the free coefficient is omitted (which is allowed), you can specify it manually here."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- finalThreshold:  Coefficient to regulate the similarity threshold. When detected, some objects can be covered by many rectangles. 0 means not to perform grouping. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def PD_default(filename):\n",
    "    image = cv2.imread(filename) # read image\n",
    "    hog = cv2.HOGDescriptor() # derive HOG features\n",
    "    hog.setSVMDetector(cv2.HOGDescriptor_getDefaultPeopleDetector()) # SVM\n",
    "    \n",
    "    # pedestrian detection\n",
    "    found, w = hog.detectMultiScale(image, hitThreshold=0, winStride=(8,8), padding=(0,0), scale=1.05, finalThreshold=2)\n",
    "    draw_detections(image, found) # draw rectangles\n",
    "    \n",
    "    # write & save image\n",
    "    cv2.imshow('original',image) # write image\n",
    "    cv2.waitKey() # for keyboard binding\n",
    "    cv2.imwrite('./img/inria.jpg',image) # save image\n",
    "    #cv2.destroyAllWindows() # clean up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "PD_default(\"./img/kids.jpg\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Daimler detector: Daimler Pedestrian Detection Benchmark Dataset is used for training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def PD_daimler(filename):\n",
    "    image = cv2.imread(filename) # read image\n",
    "    hog = cv2.HOGDescriptor((48,96), (16,16), (8,8), (8,8), 9) # derive HOG features\n",
    "    hog.setSVMDetector(cv2.HOGDescriptor_getDaimlerPeopleDetector()) # SVM\n",
    "    \n",
    "    # pedestrian detection\n",
    "    found, w = hog.detectMultiScale(image, hitThreshold=0, winStride=(8,8), padding=(32,32), scale=1.05, finalThreshold=2)\n",
    "    draw_detections(image, found) # draw rectangles\n",
    "    \n",
    "    # write & save image\n",
    "    cv2.imshow('original',image) # write image\n",
    "    cv2.waitKey() # for keyboard binding\n",
    "    cv2.imwrite('./img/daimler.jpg',image) # save image\n",
    "    #cv2.destroyAllWindows() # clean up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "PD_daimler(\"./img/kids.jpg\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  },
  "widgets": {
   "state": {},
   "version": "1.1.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}