Jocelyn Flores Villaverde

The Philippines’ mussel industry plays a vital role in coastal livelihoods and food security, yet accurate species identification remains labor-intensive. This study developed an image classification system using MobileNet on a Raspberry Pi 4 to automate the identification of Perna viridis, Modiolus Philippinarum, and Mytella charruana. Trained on 900 images and tested on 180 unseen samples, the system achieved 100% accuracy for the first two species and 85% for Mytella charruana, with an overall accuracy of 95%. Results demonstrate MobileNet’s effectiveness in fine-grained classification, offering a cost-efficient solution for aquaculture, biodiversity monitoring, and sustainable marine resource management.

Native chicken production in the Philippines is increasing, accounting for nearly half of the total population of raised chickens. Health-conscious consumers prefer native chicken due to its lower fat content. To support this growth, the government established a breeding facility featuring 10 pens, each housing 2 to 6 laying hens and a rooster, which began operation in November 2023. In recent months, staff observed a decline in laying performance in some pens. Because chicken behavior is a key indicator of growth and production performance, this study aims to implement a real-time laying hen activity recognition system using You Only Look Once Version 11 (YOLOv11) to classify hen behaviors into multiple categories. These include active behaviors (walking, eating, drinking, pecking, dust bathing, and preening), inactive behaviors (resting or inactivity), and environmental objects (feeders and water cans). A dataset of 464 images was collected from the breeding facility in Zamboanga City, Philippines. To capture hen behavior, a TP-Link Tapo C510W outdoor WiFi camera was mounted on the ceiling at a height of 80 cm above the ground. The model demonstrated excellent performance in detecting static objects such as feeders and water cans. Among behaviors, pecking and walking were identified as the most common, while drinking and dust bathing were relatively rare. The YOLOv11-based activity recognition system successfully achieved real-time classification of hen behaviors with strong performance across most activity classes. The system reached 95% mAP50, with particularly high accuracy in detecting static objects and distinctive behaviors, thereby providing a solid foundation for future improvements in recognizing more complex or challenging behaviors.

Thread counting is used to assess the quality and cultural significance of Philippine indigenous textiles such as Kalinga and Piña. We developed a portable system that automates the process using a Raspberry Pi 4 and the Alex Krizhevsky Convolutional Neural Network. The system processes textile images, employing AlexNet to count warp and weft threads, and displays results for real-time fabric assessment. Initial tests yielded an accuracy rate of ninety-six percent. By integrating AI and portability, this work provides a technical solution while contributing to the sustainability of cultural heritage.

Parakeets exhibit many similar traits across species, with only subtle differences in features and coloration used for classification, which complicates detection and identification for birdwatchers, breeders, and researchers. Traditional classification methods rely on observation, while more expensive options involve DNA sampling. We developed a bird classification system that identifies Asiatic parakeets by combining You Only Look Once Version 5 (YOLOv5) for detection with ResNet-50 for the classification of four specific species: Alexandrine, Moustache, Plum-headed, and Indian Ringneck parakeets. Using a Raspberry Pi 4B and a Raspberry Pi Camera housed in a customized enclosure to capture images of the birds, the evaluation indicated an overall accuracy of 95.05% through a multi-class confusion matrix, demonstrating the effectiveness of the system as a reliable tool for avian identification and research.

This study investigated the impact of image capture distance on the performance of convolutional neural networks (CNNs) in classifying fabrics. Unlike previous works that rely solely on digital zoom and data augmentation to simulate multi-scale variations, this research explores the use of physically captured images at far, mid-range, and near focal lengths using a camera with an attached varifocal lens. Fabric samples from three categories of Cotton, Linen, and Silk were imaged under consistent lighting to create an image dataset with a total of 1350 images used to train CNN models via transfer learning, with MobileNetV2 and ResNet50 as the baseline architectures. Classification performance was evaluated separately on each focal subset and on their combined dataset to test the trained model generalization capability. Results showed an absolute accuracy gain of 20.57% with MobileNetV2 and 9.78% for ResNet50 while performing with an improved accuracy at 98.42% for MobileNetV2 and ResNet50 at 96.30%

Cherry shrimp (Neocaridina davidi) are a popular ornamental freshwater species known for their bright colors and ability to thrive in a variety of tank environments. However, due to their small size and the subtle differences between males and females, it can be challenging to determine their sex. A You Only Look Once Version 9 (YOLOv9) object identification model and a Raspberry Pi 4-based system are used in this study to infer and classify the sex of cherry shrimp. A graphical user interface facilitates image collection and classification and displays the results. We developed a Raspberry Pi 4-based device with a camera module that captures images of cherry shrimp and integrated a DynamicDet architecture with Programmable Gradient Information and Generalized Efficient Layer Aggregation Networks to classify the sex of cherry shrimp. We evaluated the performance of the model using a confusion matrix to measure the accuracy of the sex classification. A confusion matrix was used to assess the collected data, and the system achieved an accuracy of 85.00%. The researchers suggest expanding the dataset to include more color variations, focusing on adding more robust male shrimp datasets, enabling the device to function without an enclosure, and updating the technology for faster inference.

Persistent communication barriers for the deaf and hard-of-hearing community in the Philippines are addressed in this study by developing a Filipino Sign Language Recognition (SLR) system. The system focuses on transactional signs commonly used in commercial environments such as markets and public facilities, thereby filling a gap left by existing SLR models. A vision-based approach was adopted, employing MediaPipe for landmark detection and Gated Recurrent Units for translating signs into text. To train the model, a custom dataset comprising 1065 video samples of 26 transactional signs was created, accounting for subtle variations in individual signing styles. The complete system was implemented on a Raspberry Pi 5 equipped with a webcam and touchscreen display. When evaluated on unseen data, the system achieved a recognition accuracy of 87%, demonstrating its potential for real-world applications in supporting commercial interactions for deaf and hard-of-hearing individuals.

The increasing global demand for plant-based meat alternatives, driven by concerns for environmental sustainability, animal welfare, and health, has led to a growing need for reliable food authentication methods. Animal-based and plant-based meat products are visually similar, which poses a challenge for consumers to distinguish them. We developed an electronic nose (e-nose) system with an array of MQ gas sensors (MQ-2, MQ-3, MQ-7, MQ-135, MQ-136, MQ-138), an Arduino MEGA microcontroller, and an LCD for displaying results. A fuzzy logic algorithm was implemented to process sensor data and enable decision-making through membership functions and IF-THEN rule evaluation to classify meat products as either animal meat or plant-based meat. The system performance was validated with 20 independent test samples. Determination accuracy for both categories, as well as the overall accuracy, was assessed using a confusion matrix. The findings demonstrate that the e-nose system can reliably distinguish between animal-based and plant-based meat products.

CNN with a Residual Network-50 (ResNet-50) architecture installed on a Raspberry Pi 5 is used to detect invasive aquatic plants in this study. By using object detection, the model recognizes water hyacinth, water lettuce, and water thyme and labels and bounds them accordingly. Images are taken by hand or at predetermined times, and verified detections are saved for later use. The adjusted ResNet-50 demonstrated 80.1% precision and 44.35% recall on validation, with 86.78% validation accuracy and 86.08% test accuracy. Target species from actual samples were successfully identified by the system.

We developed a prototype for grading various classes of honey. The honey was categorized into light amber, medium amber, and dark amber. The data collected were split into training and validation using an 80:20 ratio. The pre-trained model of the Mobile Convolutional Neural Network Version 2 architecture was employed. Then, the model was fine-tuned and trained, achieving the training and validation accuracy of 98.09 and 96.5%, respectively. The model was integrated into the user interface and deployed using the Raspberry Pi. The model was evaluated and yielded a 91.1% testing accuracy. Overall, the developed prototype demonstrated the feasibility of deploying it in a real-world scenario for honey grading using edge devices.