IMNets: Deep Learning Using an Incremental Modular Network Synthesis Approach for Medical Imaging Applications

Abstract

Deep learning approaches play a crucial role in computer-aided diagnosis systems to support clinical decision-making. However, developing such automated solutions is challenging due to the limited availability of annotated medical data. In this study, we proposed a novel and computationally efficient deep learning approach to leverage small data for learning generalizable and domain invariant representations in different medical imaging applications such as malaria, diabetic retinopathy, and tuberculosis. We refer to our approach as Incremental Modular Network Synthesis (IMNS), and the resulting CNNs as Incremental Modular Networks (IMNets). Our IMNS approach is to use small network modules that we call SubNets which are capable of generating salient features for a particular problem. Then, we build up ever larger and more powerful networks by combining these SubNets in different configurations. At each stage, only one new SubNet module undergoes learning updates. This reduces the computational resource requirements for training and aids in network optimization. We compare IMNets against classic and state-of-the-art deep learning architectures such as AlexNet, ResNet-50, Inception v3, DenseNet-201, and NasNet for the various experiments conducted in this study. Our proposed IMNS design leads to high average classification accuracies of 97.0%, 97.9%, and 88.6% for malaria, diabetic retinopathy, and tuberculosis, respectively. Our modular design for deep learning achieves the state-of-the-art performance in the scenarios tested. The IMNets produced here have a relatively low computational complexity compared to traditional deep learning architectures. The largest IMNet tested here has 0.95 M of the learnable parameters and 0.08 G of the floating-point multiply–add (MAdd) operations. The simpler IMNets train faster, have lower memory requirements, and process images faster than the benchmark methods tested.

Deep and Hybrid Learning Technique for Early Detection of Tuberculosis Based on X-ray Images Using Feature Fusion

“…Therefore, pooling layers solve this challenge by reducing the dimensionality of high-dimensionality deep features. Deep learning models provide two techniques for dimensionality reduction, the max pooling technique and the average pooling technique [32]. The max pooling technique selects the groups of pixels according to the size of the max pooling filter [33].…”

Integrated Design of Optimized Weighted Deep Feature Fusion Strategies for Skin Lesion Image Classification

“…Redha Ali et al [ 43 , 44 ] proposed DL-based skin lesion analysis models in 2019 and 2022. In [ 43 ], the authors proposed a CNN-based ensemble method by utilizing VGG19-UNet, DeeplabV3+, and a few other pre-processing methodologies using the ISIC 2018 challenge dataset.…”

Deep Learning for Automated Elective Lymph Node Level Segmentation for Head and Neck Cancer Radiotherapy

“…Over the last few years, developments in deep learning approaches have shown impressive results for automated segmentation of organs at risk (OAR) by using convolutional neural networks (CNN) [ 3 , 4 , 5 , 6 ] and for pathology detection [ 7 ], including the deep learning-based delineation of elective targets such as the combinations of neck LN levels, which has only more recently been investigated [ 8 ]. Most studies that demonstrated automated LN segmentation with deep learning, incorporated all LN levels or all of those levels relevant to the primary HNC location in one structure, rather than focusing on individual LN levels [ 7 , 8 , 9 , 10 , 11 ]. The methods that segment multiple lymph levels in one structure, however, are not generalizable to all primary HNC locations and tumour stages and require separate networks for contouring different combinations of lymph node levels.…”