Brain tumor detection using proper orthogonal decomposition integrated with deep learning networks.

BACKGROUND AND OBJECTIVE: The central organ of the human nervous system is the brain, which receives and sends stimuli to the various parts of the body to engage in daily activities. Uncontrolled growth of brain cells can result in tumors which affect the normal functions of healthy brain cells. An automatic reliable technique for detecting tumors is imperative to assist medical practitioners in the timely diagnosis of patients. Although machine learning models are being used, with minimal data availability to train, development of low-order based models integrated with machine learning are a tool for reliable detection. METHODS: In this study, we focus on comparing a low-order model such as proper orthogonal decomposition (POD) coupled with convolutional neural network (CNN) on 2D images from magnetic resonance imaging (MRI) scans to effectively identify brain tumors. The explainability of the coupled POD-CNN prediction output as well as the state-of-the-art pre-trained transfer learning models such as MobileNetV2, Inception-v3, ResNet101, and VGG-19 were explored. RESULTS: The results showed that CNN predicted tumors with an accuracy of 99.21% whereas POD-CNN performed better with about 1/3rd of computational time at an accuracy of 95.88%. Explainable AI with SHAP showed MobileNetV2 has better prediction in identifying the tumor boundaries. CONCLUSIONS: Integration of POD with CNN is carried for the first time to detect brain tumor detection with minimal MRI scan data. This study facilitates low-model approaches in machine learning to improve the accuracy and performance of tumor detection.

Impact of flow regime on the performance of anti-biofouling coatings.

Biofouling poses significant challenges for marine transportation due to increased skin drag, which results in increased fuel cost and associated [Formula: see text] emissions. Current antifouling methods involving polymer coating, biocides, and self-depleting layers harm marine ecosystems and contribute to marine pollution. Significant advancements have resulted in using bioinspired coatings to address this issue. However, prior investigations have predominantly focused on wettability and adhesion aspects, resulting in a limited understanding of the impact of flow regime on bioinspired structure patterns for antifouling. We conducted comprehensive experiments with two bioinspired coatings1 under laminar and turbulent flow regimes and compared them with a smooth surface. The two coatings are composed of regular arrangements of micropillars measuring 85 µm in height and spaced at 180 µm (pattern A) and 50 µm high micropillars spaced at 220 µm (pattern B). Theoretical arguments indicate that wall-normal velocity fluctuations near the micropillars' top significantly contribute to reducing the onset of biofouling under turbulence compared to the smooth surface. Pattern A coating can effectively decrease biofouling by 90% for fouling sizes exceeding 80 microns when compared to a smooth surface subjected to a turbulent flow regime. The coatings exhibited comparable anti-biofouling properties under a laminar flow. Also, the smooth surface experienced substantially higher biofouling under laminar flow compared to turbulent conditions. This underscores how the effectiveness of anti-biofouling approaches is critically influenced by the flow regime.

Novel sustainable filter for virus filtration and inactivation.

The COVID-19 pandemic has caused a multi-scale impact on the world population that started from a nano-scale respiratory virus and led to the shutdown of macro-scale economies. Direct transmission of SARS-CoV-2 (Severe Acute Respiratory Syndrome Coronavirus 2) and its variants through aerosolized droplets is a major contributor towards increasing cases of this infection. To curb the spread, one of the best engineered solutions is the use of face masks to prevent the passage of infectious saliva micro-droplets from an infected person to a healthy person. The commercially available masks are single use, passive face-piece filters. These become difficult to breathe in during strenuous activities. Also, they need to be disposed regularly due to accumulation of unwanted particulate and pathogens over time. Frequent disposal of these masks is unsustainable for the environment. In this study, we have proposed a novel design for a filter for enhanced virus filtration, better breathability, and virus inactivation over time. The filter is called Hy-Cu named after its (Hy) drophobic properties and another significant layer comprises of copper (Cu). The breathability (pressure drop across filter) of Hy-Cu is tested and compared with widely used surgical masks and KN95 masks, both experimentally and numerically. The results show that the Hy-Cu filter offers at least 10% less air resistance as compared to commercially available masks. The experimental results on virus filtration and inactivation tests using MS2 bacteriophage (a similar protein structure as SARS-CoV-2) show that the novel filter has 90% filtering efficiency and 99% virus inactivation over a period of 2 h. This makes the Hy-Cu filter reusable and a judicious substitute to the single use masks.