MAIRA-2: Grounded Radiology Report Generation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Bannur, Shruthi, Bouzid, Kenza, Castro, Daniel C., Schwaighofer, Anton, Thieme, Anja, Bond-Taylor, Sam, Ilse, Maximilian, Pérez-García, Fernando, Salvatelli, Valentina, Sharma, Harshita, Meissen, Felix, Ranjit, Mercy, Srivastav, Shaury, Gong, Julia, Codella, Noel C. F., Falck, Fabian, Oktay, Ozan, Lungren, Matthew P., Wetscherek, Maria Teodora, Alvarez-Valle, Javier, Hyland, Stephanie L.
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910615039115264
author Bannur, Shruthi
Bouzid, Kenza
Castro, Daniel C.
Schwaighofer, Anton
Thieme, Anja
Bond-Taylor, Sam
Ilse, Maximilian
Pérez-García, Fernando
Salvatelli, Valentina
Sharma, Harshita
Meissen, Felix
Ranjit, Mercy
Srivastav, Shaury
Gong, Julia
Codella, Noel C. F.
Falck, Fabian
Oktay, Ozan
Lungren, Matthew P.
Wetscherek, Maria Teodora
Alvarez-Valle, Javier
Hyland, Stephanie L.
author_facet Bannur, Shruthi
Bouzid, Kenza
Castro, Daniel C.
Schwaighofer, Anton
Thieme, Anja
Bond-Taylor, Sam
Ilse, Maximilian
Pérez-García, Fernando
Salvatelli, Valentina
Sharma, Harshita
Meissen, Felix
Ranjit, Mercy
Srivastav, Shaury
Gong, Julia
Codella, Noel C. F.
Falck, Fabian
Oktay, Ozan
Lungren, Matthew P.
Wetscherek, Maria Teodora
Alvarez-Valle, Javier
Hyland, Stephanie L.
contents Radiology reporting is a complex task requiring detailed medical image understanding and precise language generation, for which generative multimodal models offer a promising solution. However, to impact clinical practice, models must achieve a high level of both verifiable performance and utility. We augment the utility of automated report generation by incorporating localisation of individual findings on the image - a task we call grounded report generation - and enhance performance by incorporating realistic reporting context as inputs. We design a novel evaluation framework (RadFact) leveraging the logical inference capabilities of large language models (LLMs) to quantify report correctness and completeness at the level of individual sentences, while supporting the new task of grounded reporting. We develop MAIRA-2, a large radiology-specific multimodal model designed to generate chest X-ray reports with and without grounding. MAIRA-2 achieves state of the art on existing report generation benchmarks and establishes the novel task of grounded report generation.
format Preprint
id arxiv_https___arxiv_org_abs_2406_04449
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MAIRA-2: Grounded Radiology Report Generation
Bannur, Shruthi
Bouzid, Kenza
Castro, Daniel C.
Schwaighofer, Anton
Thieme, Anja
Bond-Taylor, Sam
Ilse, Maximilian
Pérez-García, Fernando
Salvatelli, Valentina
Sharma, Harshita
Meissen, Felix
Ranjit, Mercy
Srivastav, Shaury
Gong, Julia
Codella, Noel C. F.
Falck, Fabian
Oktay, Ozan
Lungren, Matthew P.
Wetscherek, Maria Teodora
Alvarez-Valle, Javier
Hyland, Stephanie L.
Computation and Language
Computer Vision and Pattern Recognition
Radiology reporting is a complex task requiring detailed medical image understanding and precise language generation, for which generative multimodal models offer a promising solution. However, to impact clinical practice, models must achieve a high level of both verifiable performance and utility. We augment the utility of automated report generation by incorporating localisation of individual findings on the image - a task we call grounded report generation - and enhance performance by incorporating realistic reporting context as inputs. We design a novel evaluation framework (RadFact) leveraging the logical inference capabilities of large language models (LLMs) to quantify report correctness and completeness at the level of individual sentences, while supporting the new task of grounded reporting. We develop MAIRA-2, a large radiology-specific multimodal model designed to generate chest X-ray reports with and without grounding. MAIRA-2 achieves state of the art on existing report generation benchmarks and establishes the novel task of grounded report generation.
title MAIRA-2: Grounded Radiology Report Generation
topic Computation and Language
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2406.04449