[W] ML/Deep learning python Dev Done

[Husires]

Hello there are a small antique store in my city want a simple website (only home page with upload bottom)

The idea of ​​the site is simple:

• The user uploads an old photo or snapshot of an antique.
• The server compares the image with the database (very small, not exceeding 500 images),
• If matching rate is more than 90% show result with information about img
• If matching rate between 70%- 90% show similar results and ask user to confirm it
• If matching rate less than 70% ask user to inter information about img and save it to database

In addition to providing a deep learning mechanism to assist admin in classifying images.

If you can do this and know someone who can do it send me the cost and time required to finish programming.
The allowed period is from 3 weeks to a month.
We can use escrow.

Done

[Joel_Jantsen]

Hello, I work in AI/ML Domain. Image classification/generating confidence score based on the similarities is something I've worked in the past with. Send me a PM to discuss further. Flexible with doing this in Python/Node whatever you prefer.

[jackg]

You'll probably need more images too just as a suggestion as 500 sounds like it'll give a low accuracy (most machine learning algorithms detect and save patterns so they'll likely not need the images once they've been passed through and described).

[Joel_Jantsen]

You'll probably need more images too just as a suggestion as 500 sounds like it'll give a low accuracy (most machine learning algorithms detect and save patterns so they'll likely not need the images once they've been passed through and described).

What OP actually wants (i think) is a classification based on 500 items that are 500 categories. So in 500 images if one of the images is a chair, our model needs to be trained to classify chairs. For that, we need to feed the model a training set comprising of 10-20 images of a chair so it classifies a newly uploaded item like a chair with a relevant confidence score. So potentially, training model based off 500 categories which is huge!