AI Paper Atlas

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

AddressCLIP: Empowering Vision-Language Models for City-wide Image Address Localization

Shixiong Xu, Chenghao Zhang, Lubin Fan, Gaofeng Meng, SHIMING XIANG, Jieping Ye

ECCV 2024 · Vision-language and multimodal › CLIP and zero-shot · approach: Recaptioning and text-supervised pretraining · 15 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 image-caption contrastive loss on top of address loss (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 77.66 -> 79.20Improved
add image-geography matching loss on top of address loss (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 77.66 -> 79.27Improved
combine caption and geography losses with address loss (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 77.66 -> 80.39Improved
train with caption loss only, without address loss (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 77.66 -> 69.27Hurt
unfreeze both image and text encoders instead of text only (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 48.88 -> 80.39Improved
unfreeze text encoder in addition to image encoder (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 77.77 -> 80.39Improved
append address text to scene caption (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 79.32 -> 80.39Improved
use BLIP-Caption-Large instead of Base for scene captions (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 80.28 -> 80.39No clear effect
reduce images per location from 24 to 3 (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 80.39 -> 56.86Hurt
train on mixed Pittsburgh + San Francisco data, test on SF (ablation)AddressCLIP · SF-IAL-BaseSSA-1 (%): 86.32 -> 85.51Hurt
restrict candidate addresses to prior streets (W=2) (ablation)AddressCLIP · Pitts-IALSSA-1 (%): 80.39 -> 89.57Improved
end-to-end AddressCLIP vs Image-GPS-Address retrieval pipelineAddressCLIP · Pitts-IALSSA-1 (%): 75.17 -> 77.01Improved

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