A python/opencv application using object detection to extract and combine objects face from images in the news in order to construct a new random avatar for Kristoffer Ørum everyday.
Facebook, twitter, tumblr & Gravatar api's
Setup a digital Ocean pyton/opencv droplet.
instruct me how to setup for subdomain redirection from main website: [url removed, login to view]
**SUGGESTED CODE STRUCTURE**
- perform every 24 hours at 10:00 CET
- Get random image search from trending <[url removed, login to view]>
- use face detection to determine if there is a face and - if not download another
- once enough images have been downloaded
- use opencv to extract from different images
- left eye
- right eye
- left eyeball
- right eyeball
- Smooth image edges
- Combine into new portrait
- resize image to fit 1920(width) x 1080(height)
- Save as both the original images from google and new image in folder label by date and accessible via the internet
- Display the new avatar at [url removed, login to view]
- submit to Facebook, tumblr, twitter and gravatar as Kristoffer Ørums new avatar.
- Well documented and structured code.
- Fully tested
- setting (frequency, image dimensions, ability to select what features detected) exposed in json file.
- Deliver a code capable of running locally on a computer, or with the instructions on how to do so.