---
description: "Essential concepts for image data curation including loading, processing, and export with GPU acceleration"
categories: ["concepts-architecture"]
tags: ["concepts", "image-curation", "tar-archives", "gpu-accelerated", "embedding", "classification"]
personas: ["data-scientist-focused", "mle-focused"]
difficulty: "beginner"
content_type: "concept"
modality: "image-only"
---

# Image Curation Concepts

This document covers the essential concepts for image data curation in NVIDIA NeMo Curator. These concepts assume basic familiarity with data science and machine learning principles.

## Core Concept Areas

Image curation in NVIDIA NeMo Curator focuses on these key areas:

<Cards>

<Card title="Data Loading" href="/about/concepts/image/data/loading">

Core concepts for loading and managing image datasets
</Card>

<Card title="Data Processing" href="/about/concepts/image/data/processing">

Concepts for embedding generation, classification, filtering, and deduplication
</Card>

<Card title="Data Export" href="/about/concepts/image/data/export">

Concepts for saving, exporting, and resharding curated image datasets
</Card>

</Cards>

## Infrastructure Components

The image curation concepts build on NVIDIA NeMo Curator's core infrastructure components, which are shared across all modalities (text, image, video). These components include:

<Cards>

<Card title="Memory Management" href="/reference/infra/memory-management">
Optimize memory usage when processing large datasets
partitioning
batching
monitoring
</Card>

<Card title="GPU Acceleration" href="/reference/infra/gpu-processing">
Leverage NVIDIA GPUs for faster data processing
cuda
dali
performance
</Card>

<Card title="Resumable Processing" href="/reference/infra/resumable-processing">
Continue interrupted operations across large datasets
checkpoints
recovery
batching
</Card>

</Cards>