Note: This is a placeholder project, that is i want to get realistic bids for a project that will not take place yet, but at a later date in time. It will be a SDK or something that can addon to pre-existing software and achieve my requirements. Please PMB for more information. If you can do it, please bid!!! and i will invite you when the real project starts!!!
This would be a three part project, using whatever open source or cheap tools you could find, but essentially i would like a toolkit that works with some demo's and also code examples to be coded into my software later on.
In all Parts the system will be a projector that projects a desktop image onto a kitchen bench (black screen unless an image is being shown), and a web camera that watches that desktop image that has been projected (which would do the recognition / tracking). It should work with different PC's and different Webcams (On Windows)
Part 1 - Finger / Hand Tracking, The app should basically work so that it can track a hand, or fingers and basically emulates a mouse, with move, click, and hold and drag functionality, so for example i have icons on my bench, i can move my hand over, and put my hand on the bench and it registers a click on that icon, i can hold down and drag that icon as another function, so for example i could click and drag an icon, or the better example is do the box feature on the Enhanced Desk video clips where your two fingers are placed on the surface and it zooms a box in between your two fingers, watch the videos to see what i mean.
Part 2 - Object Recognition, somehow the app should be able to detect items, and recognize what they are, preferably from a central remote database some how. So instead of each camera having to learn items, they can be learnt from each camera, and once one camera has learn the item all other camera's are aware of its title, description etc. In addition it should project a circle around the item, and also (with the option to turn it off) project the name of the item on top of it, the same as the Enhanced Desk project does. I am open to ideas on how the central database system may be implemented, but i would prefer it to be done using a mysql database if possible, as i have a linux dedicated server which may be of some (little use). The concept is that it could be rolled out in the kitchen and detect various implements, bowls, plates, magazines, ingredients etc and be able to recognise them without each user having to train the system to as what each item may be.
Part 3 - Augmented Images, the app should function similar to the augmented reality video in that when a certain object is detected, another augmented image can be projected over that image (given at different sizes etc) so for example say a bowl was detected, i should be able to project a still image, or better yet some sort of movie of cracking an egg, or putting something into that bowl by projecting the image over the detected object. Perhaps the best way to achieve this though is not to recognise an image and project over it, but have the system recognise an image, that image is designated a description (eg Bowl) and the Augmented system then overlays the image / video on the objects recognised as "Bowl's"
Augmented part should not look for specific symbols, eg markers but instead somehow recognise objects, and then according to software rules place image (settable size) or same for video over that item,
Open for your suggestion and your ideas, id love it if you could also send me some samples of work you may have done, or any ideas. If possible keep the budget as low as possible, whilst its cool technology it would actually be implemented into software that whilst it has commercial applications, it would be marketed as free software (advertising supported) and as such it has a low budget so bare that in mind, but bid what you think its worth and what you can :).
These are some of the videos of other projects that sort of highlight what i mean.
[login to view URL] - Enhanced Desk Video(s) Continued Below
[login to view URL]
[login to view URL]
[login to view URL]
[login to view URL] - Augmented Reality Video(s) continued below.
[login to view URL]
[login to view URL]
[login to view URL]
[login to view URL]
[login to view URL]
I have been working in this area (robotics theory, computational geometry) as a researcher since last 3 years at the Indian Institute of Science, Bangalore. I also have programming skills required for this project.
I am extremely interested.
hi this is the project of digital image processing(dip)
and the best software in the world for this is simulink and one of passion
just give me a chance to prove my expertiese in simulink/dip
I was involved in a project, closer to this - image recognition, tracking moving parts etc.. It was for Israeli Defense Forces in 2001. I was responsible for image recognition, drivers for cameras and general project management.
On a stage, you describe such a project, it can take more than 5 month with multiple milestones and a lot of recurrent tries until all your requirements are satisfied. In this case (when no clear timelines are present) I suggest working on a time basis - $25/hr for a single programmer.
Next step/next talk - google for Alexey Abraham :)
In a case, you'd like to negotiate -
Hi,
I have the experience of developing the Object Detection.I have published more than 5 papers in this area.
In additional, i am familiar with OPCV and i have the experience of porting the OPCV algorithm into embeded platform.
If you have interesting, please contact me for detail information