GLFNET: Global-Local (frequency) Filter Networks for efficient medical image segmentation

arXiv:2403.00396v1 Announce Type: new
Abstract: We propose a novel transformer-style architecture called Global-Local Filter Network (GLFNet) for medical image segmentation and demonstrate its state-of-the-art performance. We replace the self-attention mechanism with a combination of global-local filter blocks to optimize model efficiency. The global filters extract features from the whole feature map whereas the local filters are being adaptively created as 4x4 patches of the same feature map and add restricted scale information. In particular, the feature extraction takes place in …

abstract architecture art arxiv attention combination cs.ai cs.cv efficiency extract feature features filter filters global image medical network networks novel performance segmentation self-attention state style transformer type