When a new brand is registered, its image has to be different from all the others already registered. Thus, the objective of this project is to show the 5 most similar brand images from a database, comparing with the new brand to be registered.
To accomplish that, Siamese Neural Networks were used:
To run this project, please follow this steps:
- Have a database with brand images
- Put the correct paths in
DataProcessing.ipynb
and follow theInitial Instructions
- Run
DataProcessing.ipynb
- Put the correct paths in
Top5Similarities.ipynb
- Run
Top5Similarities.ipynb
In the end, the result should be similar to this:
Please read report.pdf
for more information about this project and Siamese Neural Networks.