Objective
It's my understanding that you have requested to make a project on fingerprint recognition, for some college level project. You're not in acquisition of data, any base paper, and open to my expertise and recommendation for the entire project. The BID shall be REVISED following discussion.
What can we do?
1) Have a GUI upload up to 5 or 7 images of fingerprint taken from a mobile camera
2) Process the image using various image processing technique like thinning, erosion, dilation, connected components, binarization with autimated thresholding, and store them
3) Obtain feature representation of the samples as signatures for verification
4) When a new set is uploaded, it should be able to match it with the previous set, and classify the same: authentic/fake
Data
However, the steps 1 and 2 if above can be replaced with acquisition of online data (since, they are not found without licensing, you must get that from a suitable link that I'll provide) and sharing it with me over Google Drive or Dropbox etc.
Here's, we shall use machine learning technique to learn and extract features from each sample, and use it to classify a new fingerprint, that is correctly recognize as many as possible.
Deliverables
.m files or .py files, few pages of report for understanding the same, additional files as needed
Approach
1) Select a base paper
2) Employ a method for recognition using just image processing or machine learning (any of 2, not both)
3) Compare with other similar methods
4) Create performance plots
5) Offer future work possibility
Process
As, standard procedure, we must agree upon a set of deadlines to track progress, deliverables, your feedback, and milestone payments, all well written in a brief PDF, before we begin.
So, more things can be discussed when you're online. Kindly, if proposal looks fine, leave your queries and clarifications even if I'm not online, shall revert back ASAP.
Thank you!