• Welcome to AppraisersForum.com, the premier online  community for the discussion of real estate appraisal. Register a free account to be able to post and unlock additional forums and features.

FYI - Convolution Neural Networks (CNN), ResNet, ..., Instance Segmentation, Object Detection

RCA

Elite Member
Gold Supporting Member
Joined
Jun 27, 2017
Professional Status
Certified General Appraiser
State
California
So, right now it is possible to create a Python script using ResNet-101 to process several thousand photos at a time in about 6 minutes on a Mac Studio M2 Ultra, 64G of RAM.

You train it on, say 3,000 images in about 5-6 minutes. You can keep training it with more photos to improve it. The more the better - if. you have time to classify taining photos.

In short, a well trained ResNet model should be able to tell you the brand and price of any appliances in a photo of a kitchen (and it will of course also be able to know that it is a kitchen). It will also be able give you a score, say 0-10 for condition and quality.

1. Imagine appraising a high end luxury home and by processing your photos it can create a list of brand types (ranges, ovens, sinks, cubboards, windows, doors, ....) with price. -- Instantly? Do you have any idea how many brands there are of these things?

2. Can you compete with that?

3. I'll let you think about.

4. Yes, it can be done.

5. No, I haven't gotten around to training my ResNet installation yet. It is a ton. of work to create the training data. And right now I have more urgent things to take care of. But I am working with it on the side.

But, you can be sure it will be part of future appraisal systems.


https://tinyurl.com/3359ebv7 (the best)



 
Find a Real Estate Appraiser - Enter Zip Code

Copyright © 2000-, AppraisersForum.com, All Rights Reserved
AppraisersForum.com is proudly hosted by the folks at
AppraiserSites.com
Back
Top