PixLift: Accelerating Web Browsing via AI Upscaling

2025-02-13

AI Summary

In high-cost, low-bandwidth regions, web images incur excessive data usage and slow loading times, hindering information accessibility and digital inclusion. To address this, we propose PixLift—a novel image optimization framework based on “server-side downsampling + client-side lightweight AI-based super-resolution.” Its key innovation is the first deployment of lightweight super-resolution models (ESRGAN, Real-ESRGAN, SwinIR) within browser extensions, enabling dynamic, transparent, and real-time image downscaling and reconstruction across mainstream websites. Evaluated on 71.4k real-world web pages and validated via user studies, PixLift reduces image data traffic by 62% on average while preserving perceptual quality: PSNR and SSIM remain close to original images, and 92% of users cannot discern quality degradation. By leveraging end-device computation to substitute for network bandwidth, PixLift delivers an efficient, backward-compatible, and high-fidelity web image acceleration solution for resource-constrained environments.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Accessing the internet in regions with expensive data plans and limited connectivity poses significant challenges, restricting information access and economic growth. Images, as a major contributor to webpage sizes, exacerbate this issue, despite advances in compression formats like WebP and AVIF. The continued growth of complex and curated web content, coupled with suboptimal optimization practices in many regions, has prevented meaningful reductions in web page sizes. This paper introduces PixLift, a novel solution to reduce webpage sizes by downscaling their images during transmission and leveraging AI models on user devices to upscale them. By trading computational resources for bandwidth, PixLift enables more affordable and inclusive web access. We address key challenges, including the feasibility of scaled image requests on popular websites, the implementation of PixLift as a browser extension, and its impact on user experience. Through the analysis of 71.4k webpages, evaluations of three mainstream upscaling models, and a user study, we demonstrate PixLift's ability to significantly reduce data usage without compromising image quality, fostering a more equitable internet.

Problem

• Reduces webpage sizes via AI image upscaling
• Improves web access in low-connectivity regions
• Balances data usage and image quality

Innovation

• AI upscaling reduces webpage sizes
• Browser extension implements PixLift
• Trades computation for bandwidth efficiency