Multi-Label MambaOut for Quality Assessment of Low-Field Pediatric Brain MR Images

Yueyue Zhu, Haotian Jiang, Rongqing Cai, Geng Chen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Magnetic Resonance Imaging (MRI) can be utilized to study the structure of pediatric brains non-invasively. In practice, low-field MRI scanners are widely adopted for pediatric brain imaging. However, the corresponding acquired MRI data usually suffers from severe artifacts, such as noise and motion. Therefore, an effective Quality Assessment (QA) method is essential. To this end, we design a Multi-Label MambaOut (MLMambaOut) model for the low-field pediatric brain MRI QA challenge. Specifically, we view this challenge as a multi-label classification task, utilizing four stages of gated convolution neural network blocks and ML-Decoder to finish the classification with class balance loss. Furthermore, we explore the performance of Mamba and some advanced models for this challenge. We performed extensive experiments on the challenge data, which is low-field and corrupted with seven kinds of artifacts. The results show that our MLMambaOut achieves superior classification results compared with other methods.

Original languageEnglish
Title of host publicationLow Field Pediatric Brain Magnetic Resonance Image Segmentation and Quality Assurance - 1st MICCAI Challenge, LISA 2024, Held in Conjunction with MICCAI 2024, Proceedings
EditorsNatasha Lepore, Marius George Linguraru
PublisherSpringer Science and Business Media Deutschland GmbH
Pages3-11
Number of pages9
ISBN (Print)9783031830105
DOIs
StatePublished - 2025
Event1st MICCAI Challenge on Low Field Pediatric Brain Magnetic Resonance Image Segmentation and Quality Assurance, LISA 2024, held in Conjunction with Medical Image Computing and Computer Assisted Intervention Conference, MICCAI 2024 - Marrakesh, Morocco
Duration: 10 Oct 202410 Oct 2024

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume15515 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference1st MICCAI Challenge on Low Field Pediatric Brain Magnetic Resonance Image Segmentation and Quality Assurance, LISA 2024, held in Conjunction with Medical Image Computing and Computer Assisted Intervention Conference, MICCAI 2024
Country/TerritoryMorocco
CityMarrakesh
Period10/10/2410/10/24

Keywords

  • Low-field pediatric brain
  • Multi-label classification
  • Quality assessment

Fingerprint

Dive into the research topics of 'Multi-Label MambaOut for Quality Assessment of Low-Field Pediatric Brain MR Images'. Together they form a unique fingerprint.

Cite this