EditJev improves image editing quality and speeds up processing time
EditJev: System One Routing for Image Editing Models
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Different image editing models work better or worse depending on the picture and instructions, making it hard to pick the best one. The authors created EditJev, a smart system that quickly guesses which editing model will do best for any given image and editing request. EditJev balances quality and speed, choosing the right editor for each case. Tests showed it improves image quality while making editing much faster than just using one fixed model.
What this means in practice
- •For photo editing software teams: Use EditJev to automatically select the best image editor to improve editing quality and reduce wait times for users.$Commercial implications: It enables photo editing apps to deliver higher quality edits faster by dynamically routing tasks to optimal models.
- •For digital advertising agencies: Deploy EditJev to speed up image editing workflows without sacrificing quality, improving campaign turnaround times.
Authors
Zhile Guan, Xu Jiang, Yule Duan, Gongbo Zhang, Shuyu Wang, Xiao Chen, Jian Zhang
Abstract
Instruction-based image editing models exhibit pronounced sample-specific performance variation: different models can produce substantially different results for the same source image and editing instruction, and the best-performing model often varies across requests. These models also differ substantially in inference latency, making fixed-model deployment inefficient. We therefore study image editing model routing, which dynamically selects an editor from a candidate pool given an input image and editing instruction. We introduce EditJev, a System One routing model that directly predicts sample-level preference scores over candidate image editing models in a single forward pass. A lightweight preference head on a vision-language backbone produces a fixed-dimensional score vector without autoregressive model-name generation or response parsing. To train EditJev, we construct pairwise preferences from offline candidate outputs and optimize a pairwise ranking objective to learn relative model suitability. At inference time, EditJev combines the predicted preference scores with model-level speed scores to balance editing quality and generation latency. Experiments on ImgEdit and GEdit-En show that EditJev improves aggregate quality over the strongest fixed editor, from 4.540 to 4.592 and from 8.194 to 8.255, respectively, while reducing estimated end-to-end latency by 60.8% and 31.7%. These results demonstrate the effectiveness of direct, structured model routing for efficient image editing.