Rock-Paper-Scissors Images Dataset from Kaggle
Link Dataset Rock-Paper-Scissors
This dataset contains images of hand gestures from the Rock-Paper-Scissors game. The images were captured as part of a hobby project where I developped a Rock-Paper-Scissors game using computer vision and machine learning on the Raspberry Pi (https://github.com/DrGFreeman/rps-cv).
Usage Information License CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Visibility visibility Public
Provenance Sources (https://github.com/DrGFreeman/rps-cv)
Maintainers Dataset owner drgfreeman (https://www.kaggle.com/drgfreeman)
Authors Name Julien de la Bruère-Terreault
Updates Expected update frequency Never
Last updated 2019-03-01
Date created 2019-01-19
Current version Version 2
How to predict images from Rock-Scissors-Paper Images Dataset?
* A success metric is that images can predict what images that is Rock or Scissors or Paper picture.
* Success means the numbers of predicting upper 90% accuracy from that images.
* Failure means the result of predicting no better than heuristics.
* Consider collected images from the dataset that categories Rock or Scissors or Paper pictures in the past. Assume that images will labialize that image or not.
The dataset modeling steps are:
- The dataset is divide into train sets and validation sets.
- Implement image augmentation.
- Use images data generator.
- The model pursues to use a sequential model.
- Model training does not exceed 30 minutes.
- The accuracy of the model is at least 90%.
- Can predict the images uploaded to Colab.
- The Accuracy and Validation accuracy reaches 99% and 98%. That means the models have success metrics.
- The graph titled "Validation Training and Accuracy" that could see, the distance of training data and validation is close so that overfitting does not occur.
Image Prediction Output
