Fast-and-Accurate-Human-Detection-with-HOG

• This project used OpenCV HOG people detector to build an accurate and fast enough implementation to detect people in images and videos.

Some Results

• 
• 

The Project Structure

├───input
        people1.jpg
        people2.jpg
        people3.jpg
        video1.mp4
        video2.mp4
        video3.mp4
        video4.mp4
├───outputs
│   └───frames
└───src
        hog_detector.py
        hog_detector_vid.py

• After cloning the repository, you need to create the input and outputs folder.
• You can find all the data in the input folder in the References section.

Executing the Python Files

• hog_detector.py: Execute this file from within the src folder in the terminal. This detects the people in images inside the input folder.
• hog_detector_vid.py: Execution details:
  • python hog_detector_vid.py --input ../input/video1.mp4 --output ../output/video1_slow.mp4 --speed slow: Use this execution command to run slow but accurate video detection algorithm.
  • python hog_detector_vid.py --input ../input/video1.mp4 --output ../output/video1_fast.mp4 --speed fast: Use this command to execute a little bit less accurate but fast video detection algorithm.

References

• Credits and Citations:
  • input/people1.jpg: Image by Free-Photos from Pixabay.
    • Link: https://pixabay.com/photos/urban-people-crowd-citizens-438393/.
  • input/people2.jpg: http://www.cbc.ca/natureofthings/content/images/episodes/pompeiipeople_listical.jpg.
  • input/people3.jpg: https://external-content.duckduckgo.com/iu/?u=https%3A%2F%2Ftse1.mm.bing.net%2Fth%3Fid%3DOIP.us0yaLcftx1jwMQ-tcw34gHaEU%26pid%3DApi&f=1.
  • input/video1.mp4: https://pixabay.com/videos/people-commerce-shop-busy-mall-6387/.
  • input/video2.mp4: https://pixabay.com/videos/pedestrians-road-city-cars-traffic-1023/.
  • input/video3.mp4: Video by Mihai Surdu from Pixabay.
    • Link: https://pixabay.com/videos/park-old-people-people-old-senior-6096/.
  • input/video4.mp4:
    • Link: https://www.pexels.com/video/athletes-warming-up-1585619/.
  • input/video5.mp5:
    • Link: https://www.youtube.com/watch?v=NyLF8nHIquM.