Intel Image Classification EDA

Exploratory Data Analysis of image dimensions, aspect ratios, RGB intensity, and class embeddings.

Modality 3: Image Data

1. Core EDA

Sample Images From Each Class

Train Dataset Distribution

Eval Dataset Distribution

Test Dataset Distribution

File Size Distribution

Pixel Intensity per Channel

Image Size Distribution

Image Quality Distribution

Aspect Ratio Distribution

2. Classification & Feature Clustering EDA

Class Correlation Matrix

UMAP Embedding

t-SNE Embedding

PCA Feature Projection

3. Key Insights

✓
Dataset Transition: The dataset has been updated to focus solely on Image Classification tasks. Object detection and segmentation properties have been omitted from this analysis pipeline.
✓
Feature Clustering: The T-SNE and UMAP visualizations provide insights into how well traditional features separate the different classes in lower-dimensional space.