Paper Title
"Progressive Web Application for Automated Counting and Analysis of Slipper-shaped Oyster (Crassostrea iredalei) Spat in Hatchery-Based Setting Utilizing Object Detection Algorithm towards Sustainable Oyster Management".

Abstract
The study of oyster farming in the Philippines plays a crucial role in managing coastal resources sustainably, developing rural livelihoods, and promoting food security and nutrition. Oyster farming in the country is in a stagnant situation due to a lack of technology to aid farming. Fishery hatchery researchers were still implementing the manual counting of oyster spat. This paper presents the development of an application named Progressive Web Application for Automated Counting and Analysis of Slipper-shaped (Crassostrea iredalei)Oyster Spat in Hatchery-Based Setting Utilizing Object Detection Algorithm towards Sustainable Oyster Management that is intended to help the oyster hatcheries in the administration of oyster growth and production that aims to improve the manual method. Agile Scrum Methodology was used to complete the research study. This research adopted an experimental design to develop a model for detecting and counting oyster spat. The study's primary goal was to determine and implement the most suitable model for detection among three models: You Only Look Once version 8 (YOLOv8), Single Shot Detection-MobileNet version 2 (SSD-MobileNetv2), and Faster RCNN-ResNet-50 in terms of accuracy and speed. Following a series ofevaluations, results indicated that the YOLOv8 model outperforms Faster RCNN and SSD in detecting oyster spat. YOLOv8 achieved the highest accuracy rate at 73.5%, surpassing Faster RCNN at 49.3% and SSD at 29.4%. YOLOv8 demonstrated the fastest inference time at 94.1 ms, outperforming both SSD at 304.35 ms and Faster RCNN at 248.11 ms. As a result, the YOLOv8 algorithm was implemented into the progressive web application to automatically detect oyster spat and generate valuable information.The application underwent software testing, including performance, stress, and compatibility testing to ensure that all components functioned as intended. Results showed that the application passed without issues, maintaining system stability and compatibility across most platforms and browsers. Validation testing, led by oyster experts, compared the application to the manual method using random test images. The automated counting method demonstrated an average accuracy of 90.66% and an average error of 9.33%, indicating efficient performance. The application underwent evaluation using the Technology Acceptance Model (TAM) and ISO 25010 software standards. It received a strongly acceptable rating, indicating that it simplifies the counting and monitoring of oysters, making it useful for oyster management. Keywords - Object Detection Algorithm, Oyster Spat, Progressive Web Application, Hatchery, Agile Scrum.