y0news
← Feed
Back to feed
🧠 AI Neutral

DLEBench: Evaluating Small-scale Object Editing Ability for Instruction-based Image Editing Model

arXiv – CS AI|Shibo Hong, Boxian Ai, Jun Kuang, Wei Wang, FengJiao Chen, Zhongyuan Peng, Chenhao Huang, Yixin Cao||2 views
🤖AI Summary

Researchers introduce DLEBench, the first benchmark specifically designed to evaluate instruction-based image editing models' ability to edit small-scale objects that occupy only 1%-10% of image area. Testing on 10 models revealed significant performance gaps in small object editing, highlighting a critical limitation in current AI image editing capabilities.

Key Takeaways
  • DLEBench is the first benchmark dedicated to evaluating small-scale object editing in instruction-based image editing models.
  • The benchmark contains 1889 samples across seven instruction types with target objects occupying only 1%-10% of image area.
  • Testing revealed significant performance gaps in all 10 evaluated image editing models for small object manipulation.
  • The benchmark introduces a dual-mode evaluation framework to address misalignment between AI judges and human assessments.
  • Results highlight the need for specialized benchmarks to advance precise local editing capabilities in AI systems.
Read Original →via arXiv – CS AI
Act on this with AI
Stay ahead of the market.
Connect your wallet to an AI agent. It reads balances, proposes swaps and bridges across 15 chains — you keep full control of your keys.
Connect Wallet to AI →How it works
Related Articles