A project uses AWS Rekognition to do video analysis.
- Re-implemented the front-end based on my previous implementation Smart Retailer;
Take a look at this file for the backend, follow the format and use your own customized contents.
https://github.com/denven/Video-Rekognition/blob/master/back-end/.env.example
The app will extrapolate simple business analytics from videos for the day to day operations.
- Number of customers in video
- Age, sex, emotions (physical)
- Average time in video
- Recurrences of previously analyzed people in videos
- Average time before recurrences
- Hit Upload Tab at the SiderBar
- Drag and drop mp4 files (files named as: VID_YYYYMMDD_HHMMSS.mp4 will be accepted)
- Upload file and drink a cup of coffee to wait until the analysis is done!
- The video analysis takes time, and the waiting time varies from duration, motions, persons, file sizes, resolutions of the video. It will awalys take several minutes to for a 20s video;
- You should have your AWS account configured, including users/roles/credential keys, Rekognition, S3, SQS, SNS services setup etc.
- Be aware of your bill if you stick to uploading large size and complex videos for analysis.
- Node.js
- Express.js
- Postgres
- AWS Rekognition, S3, SNS, SQS
- Server Sent Events
- React.js
- Material UI
- Material table
- reCharts