I need an expert with experience in CV, preferably OpenCV but is not mandatory as log as your code is cross-platform.
The job that the code / app has to do is to generate a 3D space / bounding box from a room's picture. The 3D space / bounding box generated will be imported into our APP that is a 3D room designer, the file the app will generate can be 3Dxml. 3DS or any other standard 3D format.
I may suppose that Pose estimation would be the best fit to achieve the goal but that is your job to tell me what would be the best approach to be used. Then make it happen, there will be 3 scenarios:
1. Already taken picture: Here we get only to process the picture and generate the 3D space / bounding box of the room.
2. Picture with Marker: We print a marker and the we take picture of the room/s that way can be calculated more accurately and we can also calculate real size values based on the printed paper in relation to the other objects.
3. Already Taken Picture + Adjust points for accuracy, some example of what I mean is seen in the video below.
(You wont be making the app just generating the 3D file from a 2d picture)
( [url removed, login to view] )
Some points to have in consideration:
1. Rooms will vary all the time
2. Pictures will be usually pictures who were already taken and from the internet
3. There may be other objects in the room
4. It wont be always a simple box as there may be more walls depending on the room configuration and angle it was taken
Only contact if you have deep understanding and experience, I am not looking to babysit or waste time if you are just trying, we need somebody who has make previous similar work or is 100% sure will deliver,
We WILL only pay AFTER Successful delivery!
No upfront payments, escrow can be used.
Provide previous experience, references of work and explain how would you achieve it in detail in order to be considered along with skype for a small interview.