Bridging the Training-Deployment Gap: Gated Encoding and Multi-Scale Refinement for Efficient Quantization-Aware Image Enhancement

2026-04-23

AI Summary

This work addresses the significant performance gap between training and deployment of image enhancement models on mobile devices, where low-precision quantization often leads to severe degradation in visual quality. To bridge this gap, the authors propose a lightweight, hierarchical neural network architecture tailored for mobile platforms, integrating gated encoding blocks with a multi-scale refinement mechanism. By incorporating quantization-aware training (QAT), the model simulates low-precision representations during training to better align with deployment conditions. This approach effectively preserves fine-grained details and substantially mitigates the quality loss typically induced by post-training quantization. As a result, the method achieves high-fidelity image enhancement while maintaining low computational overhead, making it well-suited for practical deployment on standard mobile devices.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Image enhancement models for mobile devices often struggle to balance high output quality with the fast processing speeds required by mobile hardware. While recent deep learning models can enhance low-quality mobile photos into high-quality images, their performance is often degraded when converted to lower-precision formats for actual use on mobile phones. To address this training-deployment mismatch, we propose an efficient image enhancement model designed specifically for mobile deployment. Our approach uses a hierarchical network architecture with gated encoder blocks and multiscale refinement to preserve fine-grained visual features. Moreover, we incorporate Quantization-Aware Training (QAT) to simulate the effects of low-precision representation during the training process. This allows the network to adapt and prevents the typical drop in quality seen with standard post-training quantization (PTQ). Experimental results demonstrate that the proposed method produces high-fidelity visual output while maintaining the low computational overhead needed for practical use on standard mobile devices. The code will be available at https://github.com/GenAI4E/QATIE.git.

Problem

• training-deployment gap
• image enhancement
• quantization
• mobile deployment
• low-precision representation

Innovation

• Quantization-Aware Training
• Gated Encoding
• Multi-Scale Refinement
• Mobile Image Enhancement
• Training-Deployment Gap