clothes-matcher

HackDavis 2018 to improve on Google Shopping by matching one step further.

Goal:

Aggregate all shopping products from different companies onto one site.
Allow user to look for more specific items based on selected images (which are closer to what they are looking for).

Creating a model that will successfully and accurately assess similarity between 2 images.
Get enough quality product data (professionally taken images, product links, product title)

Use NLP on product titles as another layer to assess similarity between products
Train different models (convolutional autoencoder?), different loss functions, and different arcitectures for best results.
Get more data.

Coded by Varun Ved & me. (+1 person last minute to help with front-end)

Name		Name	Last commit message	Last commit date
Latest commit History 30 Commits
Love		Love
frontend		frontend
imagesFromGoogle		imagesFromGoogle
.gitignore		.gitignore
GoogleCrawler.py		GoogleCrawler.py
README.md		README.md
csvToDb.py		csvToDb.py
depickleAndDataframe.py		depickleAndDataframe.py
fixCsvWithCorrectImages.py		fixCsvWithCorrectImages.py
getTitlesForLoveydove.py		getTitlesForLoveydove.py
google.sqlite		google.sqlite
google1937.p		google1937.p
img_to_category.ipynb		img_to_category.ipynb
main.py		main.py
plantoloveforever.txt		plantoloveforever.txt
predict.ipynb		predict.ipynb
predict.py		predict.py
progress.txt		progress.txt
resnet.ipynb		resnet.ipynb
similarity.ipynb		similarity.ipynb
similarity.py		similarity.py