Image Processing & Recognition Project

I am looking for an expert in image processing to work on an object recognition project. The initial prototype will be of limited scope leading to a longer term project upon success. There will exist a human tagged database of images and each new image must somehow be classified against and added to this database e.g. We may have 1,000 images of motorbikes and an uploaded image must be given a weighted score as to it's motorbike-likeness. Overtime the alogorithm should improve its ability to recognise motorbikes. It should also try and recognise the color of the object. So in this example it may process an image and classify it as: object:motorbike 75%, color:red 90% I imagine that a combination of processing techniques will be required to process the images in multiple steps. There can be a human powered process in place to validate the results. It is expected that some sort of machine learning algorithm that can take use of this human validation will be included in the final design. Where possible, available open source software should be used in the design e.g. openCV If this is your field of expertise and you are capable of delivering working code, please get in touch for more details. This project is for the prototype build and system design only.