3D Medical Image Analysis with Autoencoder-Based Feature Extraction and Shallow Models


Por: Tapia H.D.L., García-Ramírez J., Escalante-Ramírez B., Olveres J.

Publicada: 1 ene 2024
Resumen:
Deep neural networks automatically extract features; however, in many cases, the features extracted by the classifier are biased by the classes during the training of the model. Analyzing 3D medical images can be challenging due to the high number of channels in the images, which require long training times when using complex deep models. To address this issue, we propose a two-step approach: (i) We train an autoencoder to reconstruct the input images using some channels in the volume. As a result, we obtain a hidden representation of the images. (ii) Shallow models are then trained with the hidden representation to classify the images using an ensemble of features. To validate the proposed method, we use 3D datasets from the MedMNIST archive. Our results show that the proposed model achieves similar or even better performance than ResNet models, despite having significantly fewer parameters (approximately 14,000 parameters). © 2024 SPIE.

Filiaciones:
Tapia H.D.L.:
 Instituto de Investigaciones en Matemáticas Aplicadas y en Sistemas, Universidad Nacional Autónoma de México, Ciudad de México, Mexico

García-Ramírez J.:
 Facultad de Ingeniería, Universidad Nacional Autónoma de México, Ciudad de México, Mexico

Escalante-Ramírez B.:
 Facultad de Ingeniería, Universidad Nacional Autónoma de México, Ciudad de México, Mexico

 Centro de Estudios en Computación Avanzada, Universidad Nacional Autónoma de México, Mexico

Olveres J.:
 Facultad de Ingeniería, Universidad Nacional Autónoma de México, Ciudad de México, Mexico

 Centro de Estudios en Computación Avanzada, Universidad Nacional Autónoma de México, Mexico
ISSN: 0277786X
Editorial
SPIE-INT SOC OPTICAL ENGINEERING, 1000 20TH ST, PO BOX 10, BELLINGHAM, WA 98227-0010 USA, Estados Unidos America
Tipo de documento: Conference Paper
Volumen: 12998 Número:
Páginas:
WOS Id: 001275369100039

MÉTRICAS