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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Medical Visual Question Answering (VQA) systems play a supporting role tounderstand clinic-relevant information carried by medical images. The questionsto a medical image include two categories: close-end (such as Yes/No question)and open-end. To obtain answers, the majority of the existing medical VQAmethods relies on classification approaches, while a few works attempt to usegeneration approaches or a mixture of the two. The classification approachesare relatively simple but perform poorly on long open-end questions. To bridgethis gap, in this paper, we propose a new Transformer based framework formedical VQA (named as Q2ATransformer), which integrates the advantages of boththe classification and the generation approaches and provides a unifiedtreatment for the close-end and open-end questions. Specifically, we introducean additional Transformer decoder with a set of learnable candidate answerembeddings to query the existence of each answer class to a givenimage-question pair. Through the Transformer attention, the candidate answerembeddings interact with the fused features of the image-question pair to makethe decision. In this way, despite being a classification-based approach, ourmethod provides a mechanism to interact with the answer information forprediction like the generation-based approaches. On the other hand, byclassification, we mitigate the task difficulty by reducing the search space ofanswers. Our method achieves new state-of-the-art performance on two medicalVQA benchmarks. Especially, for the open-end questions, we achieve 79.19% onVQA-RAD and 54.85% on PathVQA, with 16.09% and 41.45% absolute improvements,respectively.

Q2ATransformer: Improving Medical VQA via an Answer Querying Decoder