Facial analytics app
£1500-4000 GBP
Paid on delivery
We need an Android app that works completely offline that returns the age and gender of a person looking into the camera. It must work offline without capturing any images or video's of the person. It must work straight from the live camera stream. The gender must be 80% accurate at least. The age should be within a 5 year range (eg. true age is 24, app returns 19 is fine). It must work pretty instantly on Android OS above 5.1 lollipop. You may assume that devices that the app will be tested on will be mid to high end devices. We have been leveraging Google Vision for facial detection, basic mood states and attention span detection. So if that particular library can be used for the initial face detection it will be more points to the developer completing this task!
Please test the app against the below databases:
[login to view URL]
[login to view URL]~vgg/software/vgg_face/
[login to view URL]
[login to view URL]
Must
1. Be an Android app
2. Work offline
2. Capture no images/video
3. Work straight from live camera feed
4. Achieve 80% accuracy for gender (On all given datasets)
5. Achieve 80% accuracy for age within a 5 year range (On all given datasets)
6. Return results within 0.2s of identifying a face (On 5.1 lollipop / mid end devices and above)
Bonus Points for
7. Utilising Google Vision library
8. Increasing age/gender accuracy or decreasing latency/'device specs needed' beyond the above criteria
Project ID: #16094531
About the project
Awarded to:
Hi\r\n\r\nFirst of all offline its not possible. \r\n\r\nWhy:\r\n\r\nBecause face detection fall in image classification. I mage classification needs machine learning algorithm and machine learning algorithm saves ther More
7 freelancers are bidding on average £3617 for this job
Hi, we've read the project description and very confident to build an Android app as per your needs. We're well-established company and have skills in all the tools and technologies, got 105+ highly skilled and experie More