Object Counting in Videos Using Python Made Easy
Ankur Shukla (~ankur67) |
Description:
With the increase in monitoring and surveillance, we have more and more video data flowing in. However sometimes it becomes difficult to automatically process this huge amount of video data even for simple tasks such as object counting due to the size and velocity of the data. All of us can appreciate applications of object counting algorithms in parking lots for calculating net available space, in public spaces for vehicle counting and or in factories for process and production monitoring.
But how easy it is to cook up a Python application for the same? Object detection can be easily done by using pre-trained models, however object counting requires to dig a little deeper to obtain satisfactory results. If you have ever tried building an object counting application in Python then you would know how difficult it is to count objects across different frames of a video. OpenCV supports several object tracking algorithms, which are fairly accurate. Yet there needs to be some work done before we can count different objects in a video using those.
This talk aims at breaking down how we can leverage OpenCV and Python to build a simple pipeline to look at video frames and easily count objects based on our requirements. During the talk we aim to look at:
- Some cool and useful object counting applications
- A naive approach for object counting and what are the bottlenecks encountered
- How can we leverage Python and OpenCV to build a robust object counting pipeline
- Try to implement our learning to build a simple vehicle counting app using Python and OpenCV
- Brief introduction to more advanced methods and concepts
- Enjoy some learning moments together :)
Prerequisites:
The talk does not require advance knowledge of concepts mentioned in the description however some prior knowledge of the following concepts would be helpful for the audience:
- Object Oriented Programming
Speaker Info:
I am a Data Scientist at Deloitte Consulting. I consult clients from different industries on their data science problems. Python is my bread and butter and I use it extensively for my day to day machine learning and data analysis tasks. I am postgraduate from CSRE, IIT Bombay in Geoinformatics and Natural Resources Engineering. Majority of my work at CSRE was focused in satellite image processing using Python.