相关实验视频

Updated: Jun 3, 2025

Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data

Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data

Published on: June 26, 2013

神经架构搜索生物医学图像分类:跨数据模式的比较研究

Zeki Kuş¹, Musa Aydin¹, Berna Kiraz²

¹Fatih Sultan Mehmet Vakif University, Department of Computer Engineering, Istanbul, Turkiye.

Artificial intelligence in medicine

|January 8, 2025

概括

两种自动化方法,PBC-NAS和BioNAS,被比较用于医学图像分类. 生物纳斯实现了更高的准确性,而PBC-NAS提供了更好的计算效率,优于现有的模型和框架.

科学领域:

人工智能的人工智能
医疗成像医学成像
计算机科学计算机科学

背景情况:

医疗图像分类的深度神经网络的手动设计耗时且不理想.
神经架构搜索 (NAS) 自动化网络设计,可能产生更高效和有效的模型.

研究的目的:

对两种新的NAS方法进行比较分析,即PBC-NAS和BioNAS.
用MedMNIST数据集评估他们在生物医学图像分类任务上的表现.
为了评估分类准确性,AUC和计算复杂性 (FLOP).

主要方法:

在MedMNIST数据集上对PBC-NAS和BioNAS进行比较分析.
评估指标:准确性 (ACC),曲线下的面积 (AUC),浮点操作计数 (FLOP).
对建筑参数,微调,搜索空间效率和差异性性能进行废除研究.

主要成果:

在准确度方面,BioNAS模型略高于PBC-NAS (BioNAS-2:0.848 avg ACC).
在PBC-NAS模型中,计算效率更高 (PBC-NAS-2:0.82 GB avg FLOPs).
这两种方法在ACC和AUC方面都超过了ResNet,AutoML框架和大多数先前的NAS研究.

结论:

PBC-NAS和BioNAS为生物医学图像分类提供高效和有效的自动化解决方案.
更大的过器尺寸和特定的模块配置提高了性能.
在没有特定数据集的NAS的情况下微调现有架构是有效的.

关键词:

生物医学图像分类的分类.美国医学会 (MedMNIST)神经架构搜索神经架构搜索基于对立的差异演变.

更多相关视频

Large-scale Reconstructions and Independent, Unbiased Clustering Based on Morphological Metrics to Classify Neurons in Selective Populations

Large-scale Reconstructions and Independent, Unbiased Clustering Based on Morphological Metrics to Classify Neurons in Selective Populations

Published on: February 15, 2017

Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies

Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies

Published on: December 15, 2023

相关概念视频

Brain Imaging

Brain Imaging

Brain imaging technologies provide critical insights into both the structure and function of the human brain, enabling medical professionals and researchers to diagnose, study, and treat neurological disorders or psychiatric disorders more effectively.
These technologies include computerized axial tomography (CAT or CT scans), positron-emission tomography (PET scans), magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI), and Transcranial Magnetic...