Search

The supply of defect-free, high-quality items is a critical factor to calculate for the long-term competitiveness of industries. Quality control is important in factories to form the item defect-free as well as to meet the needs of customers. With later advancements in deep learning and computer vision advances, it has ended up identifying different flaws from the images with near-human exactness. By introducing an insightful assembling framework, defects can be limited, and human expenses can be brought down to empower feasible development. The smart production line employs production equipment, functional testing equipment, and defect detection equipment. This paper is presented with a flaw identification system for application in smart factories based on deep learning. Training with deep learning method, EfficientDet to naturally identify deformity from ultrasonic pictures. The test results show that it was able to classify flawed items rapidly with high precision in a real-world manufacturing environment. The proposed model during 5-fold cross-validation accomplished 96% of mean accuracy was a noteworthy advancement compared to a few comparative strategies that were already utilized for this task. EfficientDet detected all the flaws present in the material during assessment and performance analysis showed it was comparatively good in detecting flaws. [ABSTRACT FROM AUTHOR]

Chronic Kidney Disease (CKD) is one amongst the main global health and well-being crisis, the lives of millions of people is claimed every year due to deprived lifestyle selection ranges and also due to genetic aspects. Hypertension and Diabetics and are the leading reasons of CKD in most countries. According to worldwide guidelines, The Glomerular Filtration Rate (GFR) and kidney damage markers are used to diagnose CKD, which is described as a disorder that causes a reduction in renal function over time. People with CKD are more likely to die early. Doctors must recognize the CKD-related disorders as soon as feasible since early identification can assist to avoid or also helps in the reversal of kidney damage. If patients are found early, they can receive better therapy and care. In several rural hospitals and health centres, there aren't enough general practitioners or nephrologists to identify the symptoms. As a result, patients have has to wait longer for a diagnosis. The researchers believe that establishing an intellectual system to categorize patients into 'Non-CKD' or 'CKD' groups will aid clinicians in dealing with several patients and providing faster diagnoses. The current research focused in analyzing the CKD prediction status using the attributes available in the dataset whereas the 25 attributes are considered inclusive of "patient ID" with 433 records. This work proposes Ensemble Ranking Support Vector Machine (ERSVM) for predicting CKD status, which considerably increases model training efficiency and achieves 0.94 high ranking accuracy. Furthermore, the suggested ERSVM approach's performance is compared to an existing method for examining the improved prediction of CKD using patient records. [ABSTRACT FROM AUTHOR]