Sauron U-Net: Simple automated redundancy elimination in medical image segmentation via filter pruning

Abstract

We introduce Sauron, a filter pruning method that eliminates redundant feature maps of convolutional neural networks (CNNs). Sauron optimizes, jointly with the loss function, a regularization term that promotes feature maps clustering at each convolutional layer by reducing the distance between feature maps. Sauron then eliminates the filters corresponding to the redundant feature maps by using automatically adjusted layer-specific thresholds. Unlike most filter pruning methods, Sauron requires minimal changes to typical neural network optimization because it prunes and optimizes CNNs jointly, which, in turn, accelerates the optimization over time. Moreover, unlike with other cluster-based approaches, the user does not need to specify the number of clusters in advance, a hyperparameter that is difficult to tune. We evaluated Sauron and five state-of-the-art filter pruning methods on four medical image segmentation tasks. This is an area where little attention has been paid to filter pruning, but where smaller CNN models are desirable for local deployment, mitigating privacy concerns associated with cloud-based solutions. Sauron was the only method that achieved a reduction in model size of over 90% without deteriorating substantially the performance. Sauron also achieved, overall, the fastest models at inference time in machines with and without GPUs. Finally, we show through experiments that the feature maps of models pruned with Sauron are highly interpretable, which is essential for medical image segmentation.

arXiv commentAccepted in Neurocomputing

References in corpus (13)

• Automated Design of Deep Learning Methods for Biomedical Image Segmentation
• Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding
• Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1
• Transfusion: Understanding Transfer Learning for Medical Imaging
• The KiTS19 Challenge Data: 300 Kidney Tumor Cases with Clinical Context, CT Semantic Segmentations, and Surgical Outcomes
• To Cluster, or Not to Cluster: An Analysis of Clusterability Methods
• An Entropy-based Pruning Method for CNN Compression
• Transfer Learning in Magnetic Resonance Brain Imaging: a Systematic Review
• Exploring Linear Relationship in Feature Map Subspace for ConvNets Compression
• RatLesNetv2: A Fully Convolutional Network for Rodent Brain Lesion Segmentation
• Structural Compression of Convolutional Neural Networks
• Deep Model Compression based on the Training History
• UCP: Uniform Channel Pruning for Deep Convolutional Neural Networks Compression and Acceleration