Bridge inspection is a crucial part of maintaining key infrastructures and ensuring public safety. However, the bridge safety inspections require substantial manpower and a variety of equipment, causing issues of high cost, accessibility limitations, and safety risks. To solve these issues, some states have begun to explore the use of rapidly growing drone technology to replace the traditional inspection method. The project teams at Marshall University created a prototype of Interactive Bridge Inspection Research using Drone (iBIRD) as proof of the concept to support the process of bridge inspection using drones. iBIRD consists of (1) a data collection module with a mobile application gathering inspection data using drones and (2) a web framework module of a database-driven Bridge Inspection Management tool. The mobile application in the data collection module is developed using React Native and the web framework of the bridge inspection management system is developed using PHP and MySQL. This paper describes details of the iBIRD system and the drone-assisted bridge inspection process. The results of this study will be of value to federal, state, and local transportation agencies and industry practitioners conducting bridge inspections. Further research will be conducted on improving data security of iBird and an automatic bridge inspection using image processing techniques with machine learning.

Periodontitis is a chronic inflammatory disease of the tissue around the teeth. The early detection of periodontitis before it manifests undesirable irreversible destruction of periodontal tissues has been an important issue in public dental health. The clinical examination is a traditional way of the diagnosis of periodontal diseases but is often insufficient and does not provide information on the current activity of periodontitis or its progression. After intensive clinical studies in the Department of Preventive and Social Dentistry at Seoul National University (SNU) in South Korea, the salivary matrix-metalloproteinase (MMP)-9 was identified as one of the major enzymes responsible for the initiation of periodontitis. SNU developed a point-of-care (POC) kit for a lateral flow test (LFT) using MMP-9 and created a diagnostic model based on a patient's personal information for screening periodontitis. After the successful clinical studies, the Rapid Periodontitis screening Tool (RPT), a database-driven web application, was developed to measure the risk of periodontitis online with the collaboration between the School of Dentistry at SNU and the Computer Science Program at Marshall University in the United States. The web interface in RPT allows anyone to enter their personal data and the value of LFT test results to receive the screening result immediately. The RPT also provides member pages to track down the test results in the long run. Once commercialized, the RPT will help early detection of periodontitis to enhance public health. This paper describes details of the tool and future research direction.

Animal behavior is an essential element in neuroscience study and noninvasive behavioral tracking of animals during experiments is crucial to many scientific pursuits. However, extracting detailed poses without markers in dynamically changing backgrounds has been a challenge. Transparent Omnidirectional Locomotion Compensator (TOLC), a tracking device, was recently developed to investigate longitudinal studies of a wide range of behavior in an unrestricted walking Drosophila without tethering and the conventional image segmentation method has been used to identify the centroids of the walking Drosophila. Since the shape or morphological features of the pixel-wise mask may vary depending on the captured images, however, the centroid calculation errors could occur when segmenting the walking Drosophila. To solve the problem, DeepLabCut, an open-source deep-learning toolbox performing markerless pose estimation on a sequence of images for quantitative behavioral analysis, was utilized to find the centroids of Drosophila melanogaster in a video recorded by TOLC. One hundred labeled images with centroids were created for the training of ResNet50 among 60,984 images and used for predicting 5,000 images in the experiment. The results of the experiment showed that the centroids predicted by the deep learning model are more accurate than the centroids from the morphological features in a specific part of the sequence of the images. Additionally, we created 200 labeled images with legs for the training of ResN et50 and predicted 5,000 images to investigate the difference between the centroids of a Drosophila melanogaster over the locations of the legs. The centroids generated from morphological features often provide incorrect information when the Drosophila melanogaster stretches out the front legs for some regions. Detailed analysis of experiment results and the future research direction with more extensive experiments are discussed.

Detecting cancer manually in whole slide images requires significant time and effort on the laborious process. Recent advances in whole slide image analysis have stimulated the growth and development of machine learning-based approaches that improve the efficiency and effectiveness in the diagnosis of cancer diseases. In this paper, we propose an unsupervised learning approach for detecting cancer in breast invasive carcinoma (BRCA) whole slide images. The proposed method is fully automated and does not require human involvement during the unsupervised learning procedure. We demonstrate the effectiveness of the proposed approach for cancer detection in BRCA and show how the machine can choose the most appropriate clusters during the unsupervised learning procedure. Moreover, we present a prototype application that enables users to select relevant groups mapping all regions related to the groups in whole slide images.

City of Huntington faces an issue of evenly distributed garbage collection among multiple fleets of trucks on different routes for fair amount of work for workers. This paper describes a method to estimate the weight of garbage picked by the trucks in each block of the city of Huntington based on a hypothesis of a correlation between the total number of stops and the total weight of garbage collected. A device was developed and mounted to the truck to collect data of truck movement. The data was analyzed to generate the number of stops the truck made and the results were compared with the weight of garbage collected throughout the route by a particular truck. Experimental results showed strong correlation between number of stops and garbage collected. Future enhancement is discussed.

Tennis ball machines help tennis players develop their game to perfect a particular stroke or area of performance. However, current commercial tennis ball machines do not provide an automatic feedback mechanism of tracking their performance, and so the collection of shot data captured during a training session is manually recorded onto paper or an electronic record. With improvements of video technology, such as computing power and resolution rate, the motion analysis and tracking of objects have been actively researched and partially successful in some sports, particularly in baseball and soccer. However, tracking a fast moving, small tennis ball is a challenging task. To solve this problem, Painless Tennis Ball Tracking System (PTBTS) has been developed to automate the collection of shot data during a tennis training session using computer vision techniques. A new algorithm has been developed to detect the boundary of court, the tennis ball's movement, and the bounce location of a player's shot. Hardware components were designed for real-time video capture and image processing as well as communication with hand-held devices. Experimental results show that our approach is robust and has a tracking accuracy that can be integrated with current commercial tennis ball machines, which would be installed by local companies. Details of the system and future enhancement are discussed.