We’re going to show how to do face recognition in the cloud using Chrysalis Cloud. We break it down for you in our video below, step-by-step, so you can easily follow along.
This demo shows all the various actions needed to build a face recognition system in Python. The steps include forming a dataset, feature detection and extraction, and finally the recognition part of the process. This example will also show you how simple it is to use Chrysalis Cloud to access and retrieve live video streamed for use in your AI, ML or CV applications. Additionally, using Chrysalis Cloud allows you to access your video without worrying about delays or frame drops.
Face recognition is a well-know and explored problem. The solutions are numerous and continuously changing. In the video we explore some of the techniques involved in face recognition using the Chrysalis Cloud Python SDK.
In the video demonstration– *scroll to end of article– we use a laptop webcam.
First, Let’s Focus in on the Chrysalis Cloud Python SDK:
Around the 10min mark, our engineer, hones in on the simple python SDK we have built. After creating the video display .py (with the proper file input packages), the simple tool, will allow you to pull the Chrysalis packages in real time to prove live stream or simply retrieve the latest image. We dive in on how to securely access the live stream, step by step.
And near the end we circle back to the exciting part, the actual face recognition. We have the video streaming to the Chrysalis Cloud from (in this case) our webcam. Next, access the video stream using Chrysalis Cloud SDK command.
The face recognition part consists of three operations.
- Detect the face
- Calculate the embeddings of the detected face
- Obtain the euclidean distance from the calculated embeddings and the embeddings from the dataset. Check if there’s a match. If so, and the match is with confidence value above a certain threshold, then label the face with the desired name.
WATCH NOW:
To follow along pull up our GitHub example and Chrysalis Cloud Tutorials:
GitHub example:
https://github.com/chryscloud/face_recognition
Chrysalis Cloud Python SDK Tutorial:
https://chryscloud.com/documentation/chrysalis-consuming-video-streams-with-python/chrysalis-cloud-python-sdk-for-computer-vision-in-the-cloud/
More tutorial about Cloud SDK usage:
https://www.youtube.com/watch?v=bWCgTW2Ar5s&feature=emb_logo&ab_channel=ChrysalisCloud
Insight about CNN:
https://insightsimaging.springeropen.com/articles/10.1007/s13244-018-0639-9
OpenCV and Python example resources:
https://www.learnopencv.com/
https://medium.com/swlh/facial-recognition-a-visual-step-by-step-d679289bab11# face_recognition
Get involved and tag us in your projects or reach out for direction to our team at: hello@chryscloud.com
Leave a Reply
You must be logged in to post a comment.