Pedestrian-detection

Implementation of the paper "Adaptive NMS: Refining Pedestrian Detection in a Crowd"

If this repository helps you in anyway, show your love ❤️ by putting a ⭐ on this project ✌️

Check out the corresponding medium blog post https://towardsdatascience.com/pedestrian-detection-using-non-maximum-suppression-b55b89cefc6.

Environment and tools

scikit-learn
scikit-Image
numpy
opencv
nms
argparse

Non Maximum Suppression

History of Oriented Gradients(HOG) combined with Support Vector Machines(SVM) have been pretty successful for detecting objects in images but the problem with those algorithms is that they detect multiple bounding boxes surrounding the objects in the image. Hence they are not applicable in our case that is detecting pedestrians on crowded roads. Here's where Non maximum suppression(NMS) comes to rescue to better refine the bounding boxes given by detectors. In this algorithm we propose additional penalties to produce more compact bounding boxes and thus become less sensitive to the threshold of NMS. The ideal solution for crowds under their pipelines with greedy NMS is to set a high threshold to preserve highly overlapped objects and predict very compact detection boxes for all instances to reduce false positives.

To execute

python run.py -i sample_images/p2.jpg

Results

References

Citing

@misc{Abhinav:2019,
  Author = {Abhinav Sagar},
  Title = {Pedestrian detection using Non Maximum Suppression},
  Year = {2019},
  Publisher = {GitHub},
  Journal = {GitHub repository},
  Howpublished = {\url{https://github.com/abhinavsagar/Pedestrian-detection}}
}

Contacts

If you want to keep updated with my latest articles and projects follow me on Medium. These are some of my contacts details:

anshitasitesh/Pedestrian-detection