English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We study multi-modal few-shot object detection (FSOD) in this paper, usingboth few-shot visual examples and class semantic information for detection,which are complementary to each other by definition. Most of the previous workson multi-modal FSOD are fine-tuning-based which are inefficient for onlineapplications. Moreover, these methods usually require expertise like classnames to extract class semantic embedding, which are hard to get for rareclasses. Our approach is motivated by the high-level conceptual similarity of(metric-based) meta-learning and prompt-based learning to learn generalizablefew-shot and zero-shot object detection models respectively withoutfine-tuning. Specifically, we combine the few-shot visual classifier and textclassifier learned via meta-learning and prompt-based learning respectively tobuild the multi-modal classifier and detection models. In addition, to fullyexploit the pre-trained language models, we propose meta-learning-basedcross-modal prompting to generate soft prompts for novel classes present infew-shot visual examples, which are then used to learn the text classifier.Knowledge distillation is introduced to learn the soft prompt generator withoutusing human prior knowledge of class names, which may not be available for rareclasses. Our insight is that the few-shot support images naturally includerelated context information and semantics of the class. We comprehensivelyevaluate the proposed multi-modal FSOD models on multiple few-shot objectdetection benchmarks, achieving promising results.

Multi-Modal Few-Shot Object Detection with Meta-Learning-Based Cross-Modal Prompting