Primary competition visual

CGIAR Wheat Growth Stage Challenge by CGIAR Platform for Big Data in Agriculture

Helping India
$3 000 USD
Completed (over 5 years ago)
Classification
Computer Vision
563 joined
203 active
Starti
Aug 28, 20
Closei
Oct 04, 20
Reveali
Oct 04, 20
No deep learning and only machine learning
Notebooks · 30 Sep 2020, 11:45 · 10

If we use only machine learning such as Light GBM model on the image pixel values read as array, we can reach <1 RMSE values. While this implementation is inferior to deep learning models, it could have some usage in resource constrained environment where availability of GPU is an issue.

Sample code here:

https://github.com/anindabitm/CGIAR-Zindi/blob/master/Tree_model.ipynb

Discussion 10 answers
User avatar
ESI SBA

Thank you, I am wondering about the training time, since you're not using any gpu and the dimensiality is so high?

30 Sep 2020, 14:14
Upvotes 0

Training time is not too high about 20-30 mins for entire pipeline

thanks that good

30 Sep 2020, 14:34
Upvotes 0

how you came up with this (7839, 3136) (7839, 3136) (7839, 3136)

30 Sep 2020, 14:42
Upvotes 0

7839 is just the size of low quality images. And 3136 is 56*56

This is pretty interesting.

30 Sep 2020, 14:55
Upvotes 0

Good work! Nice to see different approaches used :) Thanks for sharing

30 Sep 2020, 18:44
Upvotes 0
User avatar
ESI SBA

I tried it, with multiples variation, and I noticed that it fits the training data very quikly, but overfit so hard, I tried diffrent methods, and they don't seem to work good, but the technique defentley can get less than 1 rmse

1 Oct 2020, 00:14
Upvotes 0

Try attempting to use extracted image embeddings (new to Zindi and this piqued my interest, an example being the following https://www.kaggle.com/s/2543927)