PocketNet: A Smaller Neural Network for Medical Image Analysis
This enables training and deployment of medical image analysis models in resource-constrained settings, representing a novel method for a known bottleneck.
The paper tackles the problem of large and complex deep learning models in medical imaging by proposing PocketNet, a paradigm that reduces model size by throttling channel growth in CNNs, resulting in comparable performance to conventional networks while reducing parameters by orders of magnitude, using up to 90% less GPU memory, and speeding up training by up to 40%.
Medical imaging deep learning models are often large and complex, requiring specialized hardware to train and evaluate these models. To address such issues, we propose the PocketNet paradigm to reduce the size of deep learning models by throttling the growth of the number of channels in convolutional neural networks. We demonstrate that, for a range of segmentation and classification tasks, PocketNet architectures produce results comparable to that of conventional neural networks while reducing the number of parameters by multiple orders of magnitude, using up to 90% less GPU memory, and speeding up training times by up to 40%, thereby allowing such models to be trained and deployed in resource-constrained settings.