AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

VLIC: Vision-Language Models As Perceptual Judges for Human-Aligned Image Compression

Kyle Sargent, Ruiqi Gao, Philipp Henzler, Charles Herrmann, Aleksander Holynski, Li Fei-Fei, Jiajun Wu, Jason Y. Zhang

CVPR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Open-vocabulary alignment from VLMs · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Diffusion DPO post-training with VLM+LPIPS reward (ablation)VLIC · MS-COCOFD-DINO: 82.31 -> 67.83Improved
add VLM reward to LPIPS-only post-training (0.07 bpp) (ablation)VLIC · MS-COCOHuman Elo: 838 -> 858Improved
add VLM reward to LPIPS-only post-training (0.21 bpp) (ablation)VLIC · MS-COCOHuman Elo: 1103 -> 1112Improved
add VLM reward to LPIPS-only post-training (0.21 bpp) (ablation)VLIC · MS-COCOFID: 1.30 -> 1.35No clear effect
ensemble VLM reward with LPIPS (ablation)VLIC · MS-COCOFD-DINO: 67.68 -> 67.83No clear effect
ensemble VLM reward with LPIPS (ablation)VLIC · MS-COCOPSNR: 21.29 -> 21.68Improved
add VLM self-ensembling over 3 seeds (ablation)VLIC · MS-COCOFD-DINO: 68.36 -> 67.83Improved
add VLM self-ensembling over 3 seeds (ablation)VLIC · MS-COCOFID: 2.15 -> 2.31Mixed
increase number of VLM random seeds (ablation)VLM judge · BAPPS2AFC accuracy: 67% to about 71% (Figure 5)Improved
use Gemini 2.5-Flash as zero-shot judge instead of LPIPSVLM judge · BAPPS-Val2AFC accuracy: 69.56 vs 69.44No clear effect
use VLM judge instead of LPIPS on compressed imagesVLM judge · Compressed images2AFC accuracy: 92.32 vs 83.80Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy