Impact of Image Resolution on Age Estimation with DeepFace and InsightFace

2025-11-18

AI Summary

Input image resolution significantly affects the accuracy of automatic age estimation in face analysis systems, yet its impact remains poorly characterized across mainstream frameworks. Method: This study systematically investigates how resolution influences age estimation performance in DeepFace and InsightFace, evaluating both models on the IMDB-Clean dataset across seven resolutions (64×64 to 416×416) using Mean Absolute Error (MAE), Standard Deviation (SD), and Median Absolute Error (MedAE). Contribution/Results: We identify 224×224 as the optimal input resolution—deviations in either direction substantially increase estimation error. At this resolution, InsightFace achieves a MAE of 7.46 years, outperforming DeepFace (10.83 years) while exhibiting faster inference. This work provides the first systematic empirical validation of the non-monotonic relationship between input resolution and age estimation error in state-of-the-art face analysis frameworks, revealing critical inflection points in the resolution–error curve. The findings offer evidence-based guidance for optimizing image preprocessing pipelines in real-world deployment scenarios.

Technology Category

Computer Vision: Biometrics, Face, Gesture & PoseMachine Learning: Evaluation and AnalysisSearch and Optimization: Evaluation and Analysis

Application Category

Search and Retrieval-Augmented AI: Web evaluation methodologies and metricsResponsible Web: Machine-in-the-loop, human agency and autonomyWeb Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Automatic age estimation is widely used for age verification, where input images often vary considerably in resolution. This study evaluates the effect of image resolution on age estimation accuracy using DeepFace and InsightFace. A total of 1000 images from the IMDB-Clean dataset were processed in seven resolutions, resulting in 7000 test samples. Performance was evaluated using Mean Absolute Error (MAE), Standard Deviation (SD), and Median Absolute Error (MedAE). Based on this study, we conclude that input image resolution has a clear and consistent impact on the accuracy of age estimation in both DeepFace and InsightFace. Both frameworks achieve optimal performance at 224x224 pixels, with an MAE of 10.83 years (DeepFace) and 7.46 years (InsightFace). At low resolutions, MAE increases substantially, while very high resolutions also degrade accuracy. InsightFace is consistently faster than DeepFace across all resolutions.

Problem

• Evaluating image resolution impact on age estimation accuracy using DeepFace and InsightFace
• Determining optimal resolution for age estimation performance across different frameworks
• Analyzing how low and high resolutions degrade age prediction accuracy

Innovation

• Evaluated age estimation accuracy across seven image resolutions
• Identified optimal performance at 224x224 pixels resolution
• Found InsightFace consistently faster than DeepFace framework