Final Project Image Classification DBS Foundation X Dicoding

DBS Foundation Coding Camp 2024: Machine Learning Developer

This image classification project focuses on classifying images of rock, paper, and scissors gestures using machine learning techniques. The dataset consists of a total of 1312 training samples and 876 validation samples, with each class (rock, paper, scissors) having a balanced distribution. The model achieved an impressive validation accuracy of 98.86%, indicating its effectiveness in accurately classifying hand gestures. With extensive training and validation data, the model demonstrates robust performance in distinguishing between different gestures, laying the foundation for applications in gesture recognition.

Dataset yang dipakai menggunakan dataset berikut : rockpaperscissors, atau gunakan link ini pada wget command: https://github.com/dicodingacademy/assets/releases/download/release/rockpaperscissors.zip.

Requirement Project

• Dataset harus dibagi menjadi train set dan validation set.
• Ukuran validation set harus 40% dari total dataset (data training memiliki 1314 sampel, dan data validasi sebanyak 874 sampel).
• Harus mengimplementasikan augmentasi gambar.
• Menggunakan image data generator.
• Model harus menggunakan model sequential.
• Pelatihan model tidak melebihi waktu 30 menit.
• Program dikerjakan pada Google Colaboratory.
• Akurasi dari model minimal 85%.
• Dapat memprediksi gambar