An End-to-End Deep Learning Approach for Quantitative Microwave Breast Imaging in Real-Time Applications

Abstract

(1) Background: In this paper, an artificial neural network approach for effective and real-time quantitative microwave breast imaging is proposed. It proposes some numerical analyses for the optimization of the network architecture and the improvement of recovery performance and processing time in the microwave breast imaging framework, which represents a fundamental preliminary step for future diagnostic applications. (2) Methods: The methodological analysis of the proposed approach is based on two main aspects: firstly, the definition and generation of a proper database adopted for the training of the neural networks and, secondly, the design and analysis of different neural network architectures. (3) Results: The methodology was tested in noisy numerical scenarios with different values of SNR showing good robustness against noise. The results seem very promising in comparison with conventional nonlinear inverse scattering approaches from a qualitative as well as a quantitative point of view. (4) Conclusion: The use of quantitative microwave imaging and neural networks can represent a valid alternative to (or completion of) modern conventional medical imaging techniques since it is cheaper, safer, fast, and quantitative, thus suitable to assist medical decisions.

Point-of-Interest Preference Model Using an Attention Mechanism in a Convolutional Neural Network

“…In the context of data mining and analysis, deep learning models (DLs) are currently paving new avenues for POI recommendation systems. Unlike the earlier shallow neural network approaches, such as artificial neural networks (ANNs) that have been under exploration for many years, DL-based structures are characterized using a considerably amplified number of continuously linked neural layers [ 48 , 49 , 50 , 51 ]. This amplified number of layers is able to mine hidden patterns and higher-level features and can discover more difficult and hierarchical relationships.…”

Microwave Breast Sensing via Deep Learning for Tumor Spatial Localization by Probability Maps

“…[20] developed a realistic numerical two-dimensional breast phantom generator, which has been made available for this study. Following the generation of 2D phantoms, a microwave imaging dataset was created in accordance with the methodology outlined in [20]. In this process, breast slices measuring 15 × 15 square centimeters were generated.…”

“…This kind of inverse problem is well-known in the scientific literature as the "inverse scattering problem" (ISP), and finding a stable solution is not trivial due to the issues of nonlinearity and ill-posedness [9]. Several inversion algorithms can be adopted to solve the ISP under consideration [10][11][12][13][14][15][16], and among them, neural networks seem to have better image reconstruction performance compared to traditional approaches [17][18][19][20][21][22].…”

“…Existing research in the field of microwave imaging has demonstrated the potential of artificial neural networks in reconstructing realistic breast images [20,21,23]. However, the reconstruction performance of these models in terms of both spatial resolution and retrieved complex permittivity values often remain insufficient to allow for accurate differentiation between breast tumors and the surrounding fibro-glandular tissue.…”

“…This study aims to build on the work of [20,34] by developing a neural network model that generates a pixel-wise tumor probability map directly from the scattering matrix. The performance of this model will be evaluated through various quantitative and visual assessments.…”