Depthwise separable convolutional neural network (sepCNN / DSCNN)
Deep Learning; Computer Vision
noun phrase
Definition: A convolutional neural network architecture, or CNN block design, that uses depthwise separable convolutions by splitting standard convolution into depthwise and pointwise operations, thereby reducing parameter count and computational cost [Wu et al. 2025].
Example in context: “This study proposes a diagnostic method based on an improved S-transform and a Depthwise Separable Convolutional Neural Network (DSCNN).” [Wu et al. 2025]
Synonyms: DSCNN; depthwise separable CNN
Related terms: depthwise separable convolution; MobileNet-style CNN; efficient CNN