I need a full solution that will provide image recognition / image comparison for large volumes of images (vector images such as a logo) against a key reference image/s.
The goal is to detect:
a. Exact matches (regardless of image format, size or colour)
b. Partial matches (altered or modified, or where the key reference image shape is found within another image.) Also regardless of image format size or colour.
When matches or partial matches are identified, they should be given a % score.
I am looking for 2 separate project phases. These are:
1. The application to accept a feed of images and apply the recognition / comparison checking, filtering & scoring.
2. A solution for acquiring or mining relevant images from the internet to the comparison system for checking against the reference image.
If you feel that you are interested in this project and have the necessary experience and/or capability, please feel free to contact me with your time required to complete the project and your hourly or daily rate, or fixed rate for the job.
I can make this project using the principles of DSP and neural networks...