OpenCV Project Pages

OpenCV project developed in Python

Computer Vision Proof of Detection This projects set up how to use OpenCV's library to create a webcam person detection system.

List of Detectors Used

• haarcascade_eye.xml : detect eye features
• haarcascade_frontalface_default.xml : detect full face features
• haarcascade_smile.xml : detect a smile on a person

Using Spyder 3.2.3 and Anaconda Navigator

First project to explore simple facial recognition through the use of Haar Features.

This library is self-contained through the OpenCV project using its 4.1.2 packages. (As of Dec. 23, 2019, OpenCV released 4.2.0) I plan to update my computer vision detection project to OpenCV 4.2.0 in the next few iterations.

OpenCV 4.1.2 in Spyder framework or Microsoft Visual Code

Overall, the opporunity presented more complex in the establishing the environment in Anaconda. The dependencies I highlighted briefly in my blog jeremywood.ai. My initial run of my python code testing better inside Visual Code 14.0. I have since adjusted the computer vision detection lines to better accommodate a new pyhton environment that I have exported into OpenCV_v3.txt for others to reveiw. There are addition packages I have added for coding amenities and platform versatility I was able to refined it better to my workflow in Spyder.

A Note on the Python Environment for openCV project 1.0

One of the largest packages to disrupt the environment:

• MKL 2019 : Intel's Math Kernel Library "highly optimized, extensively threaded, and thread-safe library of mathematical functions for engineering, scientific, and financial applications that require maximum performan" Intel Developer Zone. I wanted it for the multithreading ability that I will soon weave into newer versions. Additionally, Intel's Deep Neural Networking code was deprecrated, and soon will be incorporated into MKL.

• Python 3.6.7 : This was my choice to launch the python platform. With a year of maturity, Oct. 2018 and its seventh maintenance update, I felt this version would have many of its bugs and nuiances understood. Additionally, for this project, the version would play a role in syntax and memory use. Later plans are for the code to update to Python 3.8.x.

• DefusedXML 0.6.0 : Knowing that I am going to use XML for the Haar Feature Detections, I wanted to add into some hardened libraries to the environment, if case I needed it. This package would be used to protect the XML imports: "XML bomb protection for Python stdlib modules"

• qtpy 1.9.0 : Versions proved with some challenges to find the one that would not break the environment. This was greatly needed for my Spyder 3.2.3 IDE version.

• pyflakes 2.1.1 : I added pyflakes for better syntax capture.

Ending the Video Detector

Press q (letter q) when the video is active

Project files:

• myface_recognition.py : Video detection of a person
• myface_recognition_smile.py : Video detection of face and a smile

---

This is an academic course project with additional changes for my customized Anaconda environment, as part of SuperDataScience instruction found on Udemy 2019