A code to train your own image dataset using Convolutional Neural Networks algorithm. The file for data can be found here : https://drive.google.com/open?id=1czBqp0Fk2F7fDoakizZPrSFRcUQ7WHow. Make sure all these dataset folders are subfolders of a main folder. The file has 531 images of rose, sunflower and tulips. The split of dataset is by using the function of sklearn. The mixed_greyscale folder is just uploaded to show how the images are resized to 200x200 and converted to grey colour. You can configure the images to red, green or blue as per the need. Again, i've tried with many layers and found this somewhat appropriate. I did try the same layer config as of for the MNIST dataset but somehow the accuracy used to get stuck between 30-40%. You can configure the layers and try getting a better accuracy. Configuring EPOCHS and BATCH_SIZE, were giving drastic changes. I'll try to add a photo of the features that the machine is classfying. Also i'll try to add confusion matrix. Feel free to ask if any doubts! Hope that helps in training a CNN model on your own pictures!
-
Notifications
You must be signed in to change notification settings - Fork 0
A code to train your own image dataset using Convolutional Neural Networks algorithm.
License
kaiwalya4850/Custom_CNN
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
A code to train your own image dataset using Convolutional Neural Networks algorithm.
Resources
License
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published