.png){kind=icon}
.png){kind=icon}
Return
TensorProjection layer: A tensor-based dimension reduction method in deep neural networks
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2025.131695.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We propose a novel dimensionality reduction method for deep neural networks, called the TensorProjection Layer (TPL). • TPL is characterized by mode-wise projection that preserves tensor structure and allows flexible control over output dimensionality. • It is applicable not only to medical image classification and segmentation but also to non-image domains such as biomedical signals and language data. • Experimental results show promising performance compared to conventional pooling methods, suggesting potential advantages in terms of generalization.
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W