This project is a POC and for:
a. match up pictures stored on computer with item catalog file.
Each Image could have more the one label. multi-label classification.
The code needs to be written in Python and run on a single windows computer and use open source or free libraries like openCV. If you will need some paid components or you believe it would do a better job please let me know in your proposal
- I have a database/CSV of products with fields: code, description and category
- Separately, I have thousands of images of products mainly fruits and vegetables in a folder (mainly one image for each unique fruits).
I need the capability of clicking on any given item in the database and it should bring up it's closest match from the images folder.
a. design a simple GUI that would display the CSV file in a grid
b. Buttons "Show Closest Match", "Show All that confidence level is above x", "Show All" and Button "Populate Full Grid" and "Clear All Grid Images"
c. Once button is pressed show matching image/s to the line
d. When showing more then one image as a result (i.e. "Show All" was clicked) I should be able to select an image and press a
button "Better Match" the system should learn and update itself
e. When ""Populate Full Grid" is clicked add image to grid and also add field showing the confidence level in the match
As an example, lets say I'm on the grid on line "Good McIntosh Apple Sale" and click show "Closed Match" and the computer gives me back more 5 kinds of apples only one is Macintosh i can now select the correct apple and press "Better Match" and the system would learn.
I would be able to add images and change the csv file and the system would keep learning
I'm new to this but assume there is already ready modals for Fruits and vegetables and we don't need learning to start from scratch
Include the word "Blue Magic" in your response and describe the technologies you would be using for this
if your offer sounds resendable I would send you link to the files