Brain tumors are the growth of abnormal cells or a mass in a brain. Numerous kinds of brain tumors were discovered, which need accurate and early detection techniques. Currently, most diagnosis and detection methods rely on the decision of neuro-specialists and radiologists to evaluate brain images, which may be time-consuming and cause human errors. This paper proposes a robust U-Net deep learning Convolutional Neural Network (CNN) model that can classify if the subject has a tumor or not based on Brain Magnetic resonance imaging (MRI) with acceptable accuracy for medical-grade application. The study built and trained the 3D U-Net CNN including encoding/decoding relationship architecture to perform the brain tumor segmentation because it requires fewer training images and provides more precise segmentation. The algorithm consists of three parts; the first part, the downsampling part, the bottleneck part, and the optimum part. The resultant semantic maps are inserted into the decoder fraction to obtain the full-resolution probability maps. The developed U-Net architecture has been applied on the MRI scan brain tumor segmentation dataset in MICCAI BraTS 2017. The results using Matlab-based toolbox indicate that the proposed architecture has been successfully evaluated and experienced for MRI datasets of brain tumor segmentation including 336 images as training data and 125 images for validation. This work demonstrated comparative performance and successful feasibility of implementing U-Net CNN architecture in an automated framework of brain tumor segmentations in Fluid-attenuated inversion recovery (FLAIR) MR Slices. The developed U-Net CNN model succeeded in performing the brain tumor segmentation task to classify the input brain images into a tumor or not based on the MRI dataset.
In the current investigation, there was the use of the technology of machine vision. The usage of this technology was informed by the need to have the laser spot’s highest energy positioned precisely, eventually allowing for the facilitation of further product work piece joining. Indeed, the joining occurred in laser welding machinery. Relative to the displacement phase, it is notable that it could aid in work piece placement into superposition areas, upon which there could be the joining of the parts. Training programs or models that were used involved convolutional neural network and deep learning, which allowed for the resultant system’s enhancement of the accuracy with which the positioning could be achieved. Also, the aforementioned algorithms were insightful because they led to the enhancement of machine work efficiency. Similarly, in the study, there was the proposing of a bi-analytic deep learning localization technique. For the purpose of system monitoring in real time, there was the use of a camera. As such, the initial stage entailed the application of the convolutional neural network, which aided in the implementation of large-scale initial searchers before having the laser light spot zone located. In turn, the phase that followed entailed increasing the camera’s optical magnification, which paved the way for the spot area’s re-imaging, as well as the application of a template matching method to ensure that high-precision repositioning was achieved. When the parameter of the search result area’s ratio was considered, it could be seen that the study was able to determine the target spot’s integrity parameters. For the case of the complete laser spot, there was the performance of the centroid calculation. Also, in situations where an incomplete laser spot reflected the target, there was the performance of invariant moments’ operation. From the findings, the study indicated that from incomplete laser spot images, the laser spot’s highest energy could be positioned precisely. The study also established that in order to establish the displacement amount, the image’s center and the laser spot’s highest energy could be overlapped.
Text recognition of images is beneficial in a wide range of computer vision purposes such as robot navigation, document analysis, and image search. The optical character recognition (OCR) technique presents a simple tool to combine text recognition functionality to many industrial and educational applications. Best OCR results can be acquired when the background of the text image is uniform and appears as a document picture. In contrast, the challenges to recognizing accurate texts occur when the image has a non-uniform background that require further preprocessing to obtain acceptable OCR result. This work discusses three scenarios. Initially, this work will test the OCR on a normal business card as an image with a uniform background. Next, discusses the text recognition of a keypad image including digits with a non-uniform background. Here, there are two preprocessing algorithms used to enhance the OCR function to overcome the negative effect of the non-uniform background of images and to detect text with high accuracy. Finally, the developed OCR method is tested on different scanned bills and discusses the variation of the obtained results. The two algorithms are the morphological reconstruction to eliminate artifacts and create cleaner images to be further processed by OCR and the Region of Interest ROI-based OCR to spot explicit regions in a tested image. Verification for the effectiveness of the Morphological-based OCR over the ROI-based method has been conducted on a dataset of scanned electricity bills images with an accuracy of 98.2 % for Morphological-based while it is only about 89.3 % for ROI-based OCR.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.