About Video Processing & Facial Expression Analysis

A tool created for one of the largest digital agencies in Japan & Singapore that work for media houses to create and manage content. It was built with video and image processing algorithms to identify any particular face from a video. The analysis further provided a clear view of the facial expression and was built using rich libraries of Python, machine learning, and data science.

Technologies Used
  • Python
  • OpenCV
  • Scikit-learn
  • Django
  • Celery
Services Provided
  • Web Application
  • Icons_g3p_v02Video & Image processing algorithms

Business Problem

The client had a requirement for a tool which can process a video and identify the facial expressions of a person. The challenge was to create an algorithm which could work frame-by-frame for the entire video. Since the videos have different sizes, aspect, ratios - it was a problem to create a tool that could work for all videos.


  1. We developed Python-based video and image processing engine to process a humongous amount of videos/images and generates a database of facial expressions.
  2. We preprocessed images/training-dataset to crop and convert images in the required format by OpenCV and generate a face expression detection dataset.
  3. The image processing and recognition took place using OpenCV and the training dataset was improved every time it processed a new image which helped to enhance the accuracy.
  4. The application decoded different file types like MP4, FLV, MVK, etc., along with the analysis of thousands of videos to ensure that the aspect, ratio, size, and quality are not compromised.

Need a similar app? But first, get your FREE QUOTE...

Recent Blogs

Stay updated with our latest blogs on Python.

Scroll to Top