Skip to content

covetedfish/Rock-Climbing-Hold-CNN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

35 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Indoor Rock Climbing Hold Classifier

A convolutional neural network for the classification of indoor rock climbing holds.

Usage

General

python3 run_nn_tf.py

Scraping the data

python3 scrape_holds.py

Processing the Images

python3 hold_identifier.py absolute/filepath/to/images

Structure

The CNN classifies images of holds into 6 categories: edges, jugs, pinches, pockets, slopers, crimps. The best accuracy I was able to achieve on the validation set was ~35% with the followin hyperparameters:

  • 8x8 pooling with stride 2
  • Two convolutional layers with ReLU activation function
    • 32 filters 5x5
    • 16 filters 3x3
  • One final dense layer with softmax activation function

The plot of training vs. validation accuracy for this model is entitled 8x8 pooling.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages