ICONIC-444: A 3.1-Million-Image Dataset for OOD Detection Research

2026-01-15

AI Summary

Existing out-of-distribution (OOD) detection research is hindered by the absence of large-scale, structured benchmark datasets with clearly defined near- and far-OOD categories, making it difficult to systematically evaluate method performance under both fine-grained and coarse-grained tasks. To address this gap, this work introduces ICONIC-444, an industrial-scale image dataset comprising 3.1 million images across 444 classes, collected from real-world sorting scenarios and featuring the first explicit distinction between near- and far-OOD categories. The dataset supports four classification and OOD detection tasks of increasing complexity and includes 22 state-of-the-art post-hoc methods as baselines. ICONIC-444 establishes a standardized, reproducible evaluation platform that fills a critical void in industrial OOD detection benchmarks and facilitates rigorous assessment of algorithmic generalization capabilities.

Technology Category

• Computer Vision: Object Detection & Categorization
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Current progress in out-of-distribution (OOD) detection is limited by the lack of large, high-quality datasets with clearly defined OOD categories across varying difficulty levels (near- to far-OOD) that support both fine- and coarse-grained computer vision tasks. To address this limitation, we introduce ICONIC-444 (Image Classification and OOD Detection with Numerous Intricate Complexities), a specialized large-scale industrial image dataset containing over 3.1 million RGB images spanning 444 classes tailored for OOD detection research. Captured with a prototype industrial sorting machine, ICONIC-444 closely mimics real-world tasks. It complements existing datasets by offering structured, diverse data suited for rigorous OOD evaluation across a spectrum of task complexities. We define four reference tasks within ICONIC-444 to benchmark and advance OOD detection research and provide baseline results for 22 state-of-the-art post-hoc OOD detection methods.

Problem

• out-of-distribution detection
• large-scale dataset
• near- to far-OOD
• computer vision tasks
• OOD benchmarking

Innovation

• out-of-distribution detection
• large-scale dataset
• industrial image classification
• near-to-far OOD
• benchmarking