Lucky High Dynamic Range Smartphone Imaging

2026-04-21

AI Summary

This work addresses the limited dynamic range of smartphone image sensors—approximately 12 stops, substantially lower than the human eye’s ~20 stops—and the shortcomings of existing HDR methods, which often produce artifacts and offer marginal improvements in handheld scenarios. The authors propose a lightweight, zero-shot HDR reconstruction approach tailored for mobile devices that generalizes across diverse smartphone cameras without requiring real-world training data. Operating in the linear RAW domain, the method processes multi-frame exposure-bracketed inputs through an iterative inference architecture based on convex combinations, representing each output pixel as a weighted average of exposure-corrected neighbors. This design effectively mitigates hallucination artifacts common in deep generative models. The framework supports an arbitrary number of input frames, robustly handles both synthetic and real unseen data, efficiently fuses 3–9 exposures, and can enhance the performance of other state-of-the-art HDR models.

Technology Category

• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Search and Optimization: Non-convex Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

While the human eye can perceive an impressive twenty stops of dynamic range, smartphone camera sensors remain limited to about twelve stops despite decades of research. A variety of high dynamic range (HDR) image capture and processing techniques have been proposed, and, in practice, they can extend the dynamic range by 3-5 stops for handheld photography. This paper proposes an approach that robustly captures dynamic range using a handheld smartphone camera and lightweight networks suitable for running on mobile devices. Our method operates indirectly on linear raw pixels in bracketed exposures. Every pixel in the final HDR image is a convex combination of input pixels in the neighborhood, adjusted for exposure, and thus avoids hallucination artifacts typical of recent deep image synthesis networks. We validate our system on both synthetic imagery and unseen real bracketed images -- we confirm zero-shot generalization of the method to smartphone camera captures. Our iterative inference architecture is capable of processing an arbitrary number of bracketed input photos, and we show examples from capture stacks containing 3--9 images. Our training process relies only on synthetic captures yet generalizes to unseen real photos from several cameras. Moreover, we show that this training scheme improves other SOTA methods over their pretrained counterparts.

Problem

• dynamic range
• smartphone imaging
• HDR
• handheld photography
• image artifacts

Innovation

• HDR imaging
• smartphone photography
• convex combination
• zero-shot generalization
• lightweight networks