We propose MedNet, a lightweight and efficient CNN architecture that integrates depthwise separable convolutions and CBAM to improve classification accuracy on medical image datasets. We design a ...
Many studies have established that the attention mechanism has great potential in improving the performance of Convolutional Neural Networks (CNNs) in image classification problems in recent years.