Article

Weakly Supervised CLIP Outperforms a 7 B Open-Source Vision–Language Model on Bi-Temporal Satellite Change Classification

2026-06-12

Abstract excerpt

<title>Abstract</title> <p>This study investigates whether a current open-source vision–language model (VLM) at the 7 B-parameter scale can classify bi-temporal change on satellite imagery zero-shot. A controlled comparison is set up on LEVIR-CC, treating its five human captions per pair as a free source of binary class labels (no-change vs. newly- completed construction) via a transparent majority-vote rule. On...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 8579bdbb-83b8-56be-98b5-92b719f7f499
DOI: 10.21203/rs.3.rs-9997769/v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Weakly Supervised CLIP Outperforms a 7 B Open-Source Vision–Language Model on Bi-Temporal Satellite Change ClassificationDOI 10.21203/rs.3.rs-9997769/v1Read

Mapping Cultural Ecosystem Service Flows from Social Media Imagery with Vision–Language Models: A Zero-Shot CLIP FrameworkSemantic proximity Read

Advancing Multimodal Visual Analytics for Disaster Management with LLM-Enhanced CLIPSemantic proximity Read

BMRS: Bongard–Maximov Problems for Remote SensingSemantic proximity Read

RobustVLM: Enhancing Large Vision-Language Models through Adaptive Visual Context Learning, Weak-to-Strong Robust Transfer, and Domain-Adaptive FeedbackSemantic proximity Read

HEADHUNTER: Training-Free Annotated Dataset Synthesis via Self-Guided Diffusion Transformer Attention Head SelectionSemantic proximity Read

Tarsier2: Advancing Large Vision-Language Models from Detailed Video Descriptions to Comprehensive Video UnderstandingSemantic proximity Read

IUNet-UCD: Improved U-Net with weighted binary cross-entropy loss function for urban change detection of Sentinel-2 satellite imagesSemantic proximity Read

mIT-CMCA: A Cross-Modal Category Alignment Framework for Robust Maize Disease IdentificationSemantic proximity Read

Fine-Grained Multimodal Alignment and Iterative Rectification Learning FrameworkSemantic proximity Read

A Review of Multimodal Vision–Language Models: Foundations, Applications, and Future DirectionsSemantic proximity Read

Research on YOLOv13n-Based Object Detection for Huashan Rock Art with Multi-Scale Optimized TrainingSemantic proximity Read