This repository contains examples and implementations of semantic image clustering using various machine learning algorithms. The main focus is on clustering images based on their semantic content, leveraging deep learning techniques.
Prerequisites
- Python 3.x
- TensorFlow
- Keras
- NumPy
- scikit-learn
- OpenCV (if needed for image preprocessing)
Image Clustering 1: This folder contains a Keras example of semantic image clustering using the SCAN (Self-labeling via Contrastive Neighbor) algorithm. The example demonstrates how to perform unsupervised clustering on a standard image dataset. -Navigate to the Image Clustering 1 folder. -Follow the instructions in the README.md file within the folder to run the SCAN algorithm example using Keras.
Image Clustering Own Dataset: In this folder, we apply the SCAN algorithm to a different dataset. This example shows how to adapt the SCAN algorithm to new datasets, providing insights into preprocessing, training, and evaluating the clustering performance on custom data. -Navigate to the Image Clustering Own Dataset folder. -Follow the steps in the README.md file to preprocess your dataset and run the clustering algorithm.
FinalSIC: This folder contains the implementation of semantic image clustering on the IEMOCAP Dataset. This example goes through the process of clustering emotional expression images, highlighting the steps involved in handling a more complex and diverse dataset. -Navigate to the FinalSIC folder. -Refer to the README.md file for detailed instructions on handling the IEMOCAP Dataset and performing semantic image clustering.