CLIP-IT: CLIP-based Pairing of Histology Images with Privileged Textual Information

Keyphrases

• Adaptation 50%
• Annotation Cost 50%
• Classification Accuracy 50%
• CLIP Model 50%
• Complementary Diagnostic Tests 50%
• Complementary Modalities 50%
• Histological Images 100%
• Image-text Pairs 50%
• Inference Time 50%
• IT Use 50%
• LoRa 50%
• Low Overhead 50%
• Medical Imaging 50%
• Multimodal Learning 50%
• Multimodal Pre-training 50%
• Paired Data 100%
• Pathology Report 50%
• Relevant Content 50%
• Semantic Gap 50%
• Text Report 100%
• Text-dependent 50%
• Textual Information 100%
• Time Complexity 50%
• Tissue Type 50%
• Unaligned 50%
• Unpaired Text 50%
• Vision Model 100%
• Vision-Language Models 50%

Computer Science

• Annotation 100%
• Classification Accuracy 50%
• Medical Imaging 50%
• Multimodal Learning 50%
• Pre-Trained Model 50%
• semantic gap 50%
• Textual Information 100%
• Time Complexity 50%
• Vision-Language Model 50%