Alzheimer's disease, the most prevalent form of dementia worldwide, progressively impairs memory and cognitive functions, ultimately leading to death. This neurodegenerative disorder affects areas of the brain responsible for thought, memory, and language, and progresses through three primary stages: mild, moderate, and severe. Currently, there is no cure for Alzheimer's; however, existing treatments can alleviate symptoms. Early detection is crucial, and Magnetic Resonance Imaging (MRI) is a fundamental diagnostic tool. MRI uses magnetic fields and radio waves to create detailed images of the brain, aiding in the assessment of Alzheimer's progression. Unfortunately, the interpretation of MRI images can be error-prone and time-consuming. This paper introduces a Convo-lutional Neural Network (CNN) model that employs artificial intelligence to analyze MRI scans, aiming to provide accurate and expedited diagnoses of Alzheimer's disease.

We proposes a novel method for camera-to-LIDAR calibration, which estimates the relative transform from LIDAR to camera sensors during on-field run-time without specialized equipment. The proposed method trains a transformer model to directly predict the 6 D pose between the camera and LIDAR. Our method uses vision transformers to extract features from the LIDAR and camera image and calculate the similarity for candidate pose hypothesis. The results demonstrate the superiority of the proposed method over baseline approaches, showcasing its potential for practical applications.

Diabetic retinopathy (DR) is a sight-threatening condition associated with diabetes, characterized by damage to the retinal blood vessels. Key to the automation of DR staging is the identification of various symptoms directly or closely associated with retinal blood vessels, as well as the number of these symptoms in the four quadrants of the retina separated by the optic disc. Therefore, precise identification of the optic disc (OD) and blood vessels in fundus images is crucial for DR stage diagnosis but is often time-consuming and requires expert analysis. This study introduces a thresholding-based approach for the automated localization of the OD and the detection of blood vessels in fundus images of diabetic patients. Our algorithm is more robust than some deep learning-based algorithms, achieving more accurate results, particularly in advanced DR stages where the resemblance between various symptoms and blood vessels complicates the extraction of blood vessels. Additionally, our computer vision system can achieve OD localization and blood vessel segmentation in real time. The experimental results on a dataset selected by an ophthalmologist from a Kaggle dataset, ensuring data quality, show that the proposed algorithm can achieve an accuracy higher than 94% for both OD localization and blood vessel detection, outperforming some state-of-the-art algorithms.

Effective communication and scheduling are essential for a productive academic environment. This paper proposes a mobile application called ProSchedule, designed to facilitate seamless interactions between students and professors at Virginia State University (VSU). The app aims to improve time management by providing a platform for students to view professors' teaching schedules, office hours, and research work. ProSchedule incorporates user-friendly design elements and features such as calendar integration, collaboration tools, and customization options. Built using Swift, SwiftUI, and Xcode, the app leverages WebKit for web content integration and tables/lists for data display. Future work will focus on enhancing user interaction, authentication features, and image integration with Firebase for personalization.

We propose a novel hybrid algorithm framework to address the problem of clustering images received in spam emails based on authorship. The multimodal nature of these images, containing foreground objects, text, or a combination of both, poses a significant challenge for grouping them effectively. To address this challenge, we train convolutional autoencoders (CAE) to extract visual features from the images, which are produced by the encoder of the trained CAEs. Furthermore, we utilize an optical character recognition (OCR) algorithm to extract text information from the images. The extracted text and visual features, in conjunction with layout features, are employed to construct matrices that measure the similarities between each pair of images in our experiment dataset. We subsequently apply a two-stage hierarchical clustering algorithm to cluster the images into groups. We compare the results produced by our proposed algorithm with the ground truth collected by a domain expert. Our experimental findings reveal that our relatively simple CAEs, with as few as thirty-seven visual features, can achieve homogeneity, completeness, and V-measures that are as high as those obtained from more complex convolutional neural networks (CNNs).

Diabetes Mellitus (DM) is a disease that affects millions of individuals globally. Diabetic Retinopathy (DR) is one of the many symptoms associated with Diabetes Mellitus. Diabetic Retinopathy occurs when chronic high glucose levels damage blood vessels within the eye, causing blood vessels to leak and in later stages cause Vitreous Hemorrhage. The key to preventing and treating advanced stages of Diabetic Retinopathy is early diagnosis. This paper introduces a method of training and boosting Deep Learning models for segmentation of Vitreous Hemorrhage in fundoscopic images, which would further facilitate classification of DR stages. The proposed algorithm generates a mask of Vitreous Hemorrhage by deploying pixel-wise binary classification to the fundoscopic images.

Metal additive manufacturing is an important manufacturing technique due to its cost effective and rapid prototyping capabilities. This technique has been widely used in various industries ranging from aerospace to military defense. Depending on different metal additive manufacturing settings, microstructures, such as cracks, would be generated for finished parts within additive manufacturing procedures. Those microstructures could lead to undesired defects, which may negatively impact the quality of the fabricated products, especially when they are delivered for mission critical tasks. This study developed a deep learning based computer vision approach for microstructure, especially crack recognition in images collected from Scanning Electron Microscope (SEM). Through performing segmentation of cracks in the SEM images of different magnifying factors, manufacturing quality could be administered quantitatively.

We investigate a deep learning-based method to estimate 6D pose information of target objects from one or multiple images. We use a modified YOLO2 as the backbone network for feature extraction and a detection network to detect the target's 3D bounding box. The network is trained using 3D errors in addition to the usual 2D pixel errors. Our method also uses a perspective-aware method to select the best keypoints to estimate the 6D pose. Synthesis data experiments show significant angular accuracy improvement in the estimated 6D pose.

We propose an algorithm framework to automatically identify the topcoat layer (TCL) in thermal barrier coating (TBC) images obtained from three different commonly used coating materials. The proposed algorithm framework utilizes a combination of image analysis techniques and statistics to identify the upper and lower boundaries of TCL, which in turn will be used to generate a TCL mask for each TBC image. To evaluate the performance of the proposed framework, we compare the TCL mask generated by the proposed framework with the ground truth on an experimental dataset that consists of a total of 157 TBC images with a roughly even distribution from three coating materials. Our proposed algorithm using a mixed method to identify the upper and lower boundaries of TCL achieves a minimum of 98% average classification accuracy and F 1 -score for all three types of the TBC images in our experimental dataset. These metrics are comparable or much higher than those obtained from using the thresholding-based algorithm framework in our previous work.

Thermal barrier coating, a widely used advanced manufacturing technique in various industries, provides thermal insulation and surface protection to a substrate by spraying melted coating materials on to the surface of the substrate. As the melted coating materials solidify, it creates microstructures that affect the coating quality. An important coating quality assessment metric that determines its effectiveness is porosity, the quantity of microstructures within the coating. In this article, we aim to build a novel algorithm to determine the microstructures in a thermal barrier coating, which is used to calculate porosity. The hybrid approach combines the efficiency of thresholding-based techniques and the accuracy of convolutional neural network (CNN) based techniques to perform a binary semantic segmentation. We evaluate the performance of the proposed hybrid approach on coating images generated from two different types of coating powders. These images exhibit various texture features. The experimental results show that the proposed hybrid approach outperforms the thresholding-based approach and the CNN-based approach in terms of accuracy on both types of images. In addition, the time complexity of the hybrid approach is also greatly optimized compared to the CNN-based approach.

This paper introduces a multimodal framework for clustering spam images received in unsolicited emails. Spam images in the same cluster have similar visual and textual contents and could be generated by a common spam source. To perform the clustering task, we first extract three main categories of features: 1) Visual features, extracted by pretrained convolutional neural networks (CNNs); 2) Layout features, the location of illustrations in the spam images; 3) Text features extracted by optical character recognition (OCR) algorithm. We then use a two-stage hierarchical clustering framework to form clusters based on the pair-wise similarity matrices of the extracted features. We evaluate the performance of the proposed approach on a 2,100 spam image dataset collected from three months of emails. The experimental results show that the proposed method achieved satisfactory clustering outcomes in terms of an external entropy-based metric, the V-measure.

Diabetes Mellitus (DM) is a health issue which is characterized by an elevated glucose level in the bloodstream. Diabetic Retinopathy (DR) stage classification plays an important role in diagnosis of patient for Diabetes Mellitus (DM). Different DR stages could normally be distinguished by various retinal lesions. In this work, Convolutional Neural Network (CNN) based retinal lesion detection method has been deployed to recognize microaneurysms (MAs) and hard exudates (HEs) in fundoscopic images in order to further facilitate DR stage classification. The proposed approach has achieved significant accuracy performance increase compared to the previous methods in DR stage classification.

Precision landing is critical for rapid launching and recovering of unmanned aerial vehicles in ship-based scenarios and space-limited environments. The solution presented here uses a deep convolutional network based vision pipeline to detect the landing mark and estimate system state in real-time. The landing sequence is controlled by vehicle trajectory points using a linear predictive algorithm. The trajectory points are fed to a low level position controller and velocity controller. Both controllers demonstrate satisfactory performance and achieve < 0.3-meter accuracy with the platform moving up to 80% of maximum UAV velocity.

This work trained convolutional neural networks (CNNs) to identify microstructure characteristics and then provide a measurement of porosity in the topcoat layer (TCL) of thermal barrier coatings using digital images captured by an inverted optical microscope. Porosity in a coating is related to thermal compensation and the longevity of the parts protected by the coating. The approach employs pixel-wise classification and transfer learning accompanied by data augmentation to expedite the training process and increase classification accuracy. The authors evaluate CNN-based models globally on the entire TCL of 159 high resolution raw images of three types (Type A, B, C) that are generated from three different types of powders and exhibit different physical and visual properties. The experimental results show that the CNN-based models outperform adaptive local thresholding-based porosity measurement (ALTPM) approach that this paper proposed in the previous work by 7.76%, 10.82%, and 12.10% respectively for Type A, Type B, and Type C images in terms of the average classification accuracy.

Diabetes mellitus (DM) is a worldwide major medical problem. Diabetic retinopathy (DR) staging is important for the estimation of DM and the evaluation of associated retinopathy. According to the international clinical diabetic retinopathy & diabetic macular edema disease severity scales, most of the dilated ophthalmoscopy observable findings are associated with retinal blood vessels. In order to objectively and accurately determine the diabetic retinopathy stages, it is essential to automatically detect and extract retinal blood vessels in fundoscopic images. This paper introduces and compares various convolutional neural networks to recognize retinal blood vessels in fundoscopic images. The experimental results demonstrate the effectiveness of the proposed approach.

The goal for this panel is to propose a schema for the advancement of intelligent systems through the use of symbolic and/or neural AI and data science.Specifically, discussants will explore how conventional numerical analysis and other techniques can leverage symbolic and/or neural AI to yield more capable intelligent systems.This approach could yield significant improvements in such domains as Meteorological and Oceanographic (METOC) signal processing, logistics, scheduling, pattern recognition, optimization, ergonomics, explanation, causal inference and prediction, system diagnostics, education and training, and a plethora of additional applications.Self-reference is inherent to autonomous thought; and, this appears to be indistinguishable from consciousness from a computability perspective.Thus, the question arises, can we program more efficient ways to support the programming (problem-solving) process?The panel will explore these and other advanced topics related to information reuse and integration and of fundamental importance to data science.

Thermal barrier coating (TBC), a widely used advanced manufacturing technique in various industries, provides thermal insulation and surface protection to a substrate by spraying melted coating materials on to the surface of the substrate. This article is an extended version of a previously published work. To quantify microstructures in the TBC, the authors introduce a fully automated image analysis-based TBC porosity measure (TBCPM) framework which includes 1) top coat layer (TCL) detection module, and 2) microstructure recognition and porosity measure module. The first module is designed to automatically identify the TCL in a TBC image using a histogram-based approach. The second module recognizes the microstructures in the TCL using a local thresholding-based method. This article extends the previous work by introducing convolutional neural networks (CNNs) to enhance the performance of the second module. The experimental results show that the CNN-based methods outperform local thresholding-based methods, and results of the proposed porosity measure are comparable to that of the domain experts.

Diabetic Retinopathy (DR) stage classification has been regarded as a critical step for evaluation and management of diabetes retinopathy. Because of damages of the retina blood vessels caused by the high blood glucose level, different extent of microstructures, such as micro-anuerysms, hard exudates, and neovascularization, could occupy the retina area. Deep learning based Convolutional Neural Network (CNN) has recently been proved a promising approach in biomedical image analysis. In this work, representative Diabetic Retinopathy (DR) images have been aggregated into five categories according to the expertise of ophthalmologist. A group of deep Convolutional Neural Network methods have been employed for DR stage classification. State-of-the-art accuracy result has been achieved by InceptionNet V3, which demonstrates the effectiveness of utilizing deep Convolutional Neural Networks for DR image recognition.

We present a vision-based search and rescue system which uses a unmanned aerial vehicle(UAV) swarm to search and track missing personnel/animals. A major benefit of multiple-vehicle search operation is the extended coverage due to the "bridging" effect between the vehicles, which allow a larger and further search area that is beyond the reach of a single vehicle. The challenge here is to plan the UAV swarm's motion paths while maintain the communication links between vehicles during the flight. Our path planning method uses a two-tie search algorithm to approximate the optimum paths for n-UAV search. The integrated vision pipeline and target recognition subsystem is evaluated with emulated UAVs and image sensors.