morganm94 / object-detection

AI-powered Tracking, Speed Monitoring, and License Plate Recognition. This project focuses on the study and implementation of various object detection techniques. It covers a wide range of objects including faces, vehicles, and more. Different methods are employed to achieve accurate and efficient detection.

Object Detection Project

Overview

This project focuses on the study and implementation of various object detection techniques. It covers a wide range of objects including faces, vehicles, and more. Different methods are employed to achieve accurate and efficient detection.

Techniques Used

The following object detection techniques have been explored in this project:

1. Face Detection

• Description: Face detection is a crucial component in many applications. This technique involves locating and identifying human faces within an image or video stream.

• Face Detection with dlib

  This project contains Python scripts for face detection using the dlib library with two

  different techniques: Histogram of Oriented Gradients (HOG) and Convolutional Neural Network (CNN).

  The scripts are designed to work with both images and videos, providing a flexible solution for face detection tasks.

The face detection scripts in this project leverage two distinct techniques for accurate and efficient face detection:

1. Histogram of Oriented Gradients (HOG)

   • Utilizes the HOG + SVM face detection algorithm provided by dlib.
   • Well-suited for real-time face detection in images and videos.
   • Provides a good balance between accuracy and speed.

   output

   • image

   

   • video

   

2. Convolutional Neural Network (CNN)

   • Uses the dlib CNN face detection model for improved accuracy.
   • Particularly effective when high accuracy is required, but it may be slower than HOG.
   • Suitable for scenarios where precision is crucial, such as image analysis.

   output

   • image

   

   • video

   

• Features

  • HOG-Based Face Detection: Efficient and real-time face detection using the HOG technique.

  • CNN-Based Face Detection: Improved accuracy through the use of a Convolutional Neural Network.

  • Image Detection: Detect faces in images and save the results.

  • Video Detection: Perform real-time face detection in videos and save the processed video.

2. vehicle Detection

• Description: vehicle detection is essential in scenarios like autonomous driving and traffic monitoring. This technique aims to identify and locate vehicles in images or video frames.

• Implementation:

  Techniques Used

  The following vehicle detection techniques have been explored in this project:

  1. YOLO (You Only Look Once)

  • Description: YOLO is a popular real-time object detection system that processes images in a single pass. It is known for its speed and accuracy in detecting objects, including vehicles.

  • Implementation: YOLO vehicle Detection Code

    • result

    

  2. Haar Cascade Classifier

  • Description: Haar Cascade Classifier is a machine learning object detection method used to identify objects in images or video streams. It can be trained to detect specific objects, including vehicles.

  • Implementation: Code

    • result