Jaundice, a common condition in infants that results from high bilirubin levels in the blood, often requires early diagnosis and monitoring to prevent severe complications, especially in newborns. Traditional diagnostic methods can be time-consuming and subject to human error. This study proposes an approach for real-time jaundice detection using advanced image processing techniques and machine learning algorithms. By analyzing images captured in RGB color spaces, pixel values are extracted and processed through Otsu’s thresholding and morphological operations to detect color patterns indicative of jaundice. A classifier model is then trained to distinguish between normal and jaundiced conditions, offering an automated, accurate, and efficient diagnostic tool. The system’s potential to operate in real-time makes it particularly suited for clinical settings, providing healthcare professionals with timely insights to improve patient outcomes. The proposed method represents a significant innovation in healthcare, combining artificial intelligence and medical imaging to enhance the early detection and management of jaundice, reducing reliance on manual interventions and improving overall healthcare delivery.
โรคดีซ่าน ซึ่งเป็นภาวะทางการแพทย์ทั่วไปที่มีลักษณะการเหลืองของผิวหนังและดวงตา มักบ่งบอกถึงความผิดปกติของตับหรือเลือดที่อยู่เบื้องหลัง การตรวจพบในระยะเริ่มต้นมีความสำคัญอย่างยิ่ง โดยเฉพาะในทารกแรกเกิด ที่หากไม่ได้รับการรักษาโรคดีซ่าน อาจนำไปสู่ภาวะแทรกซ้อนร้ายแรงได้ วิธีการวินิจฉัยแบบดั้งเดิมต้องอาศัยการตรวจสอบด้วยสายตาหรือการทดสอบในห้องปฏิบัติการ ซึ่งอาจใช้เวลานานและมีข้อผิดพลาดได้ ความก้าวหน้าล่าสุดในด้านการประมวลผลภาพและแมชชีนเลิร์นนิงเสนอความเป็นไปได้ใหม่ ๆ สำหรับการตรวจจับที่แม่นยำ มีประสิทธิภาพ และแบบเรียลไทม์มากขึ้น ด้วยการวิเคราะห์รูปแบบสีผิว ปัญญาประดิษฐ์ (AI) สามารถทำให้การวินิจฉัยเป็นไปโดยอัตโนมัติ ทำให้รวดเร็วขึ้นและลดการพึ่งพาการประเมินโดยมนุษย์
คณะวิทยาศาสตร์
Albumin Smart Test is an innovative device for screening of kidney disease by mobile phone. The device composes of (1) container and testing device with specific reagents for the albumin detection. (2) The mobile phone, installed with "Albumin smart test" application. The test is started by dropping patient urine and the reagent. Color of the product is occurred and is captured by the application with subsequent evaluation of the albumin amount. The results is displayed on screen within 3 mins. This innovative device is simple, rapid and user-friendly.
คณะแพทยศาสตร์
This study explores the application of deep convolutional neural networks (CNNs) for accurate pill identification, addressing the limitations of traditional human-based methods. Using a dataset of 1,250 images across 10 household remedy drugs, various CNN architectures, including YOLO models, were tested under different conditions. Results showed that natural lighting was optimal for imprinted pills, while a lightbox improved detection for plain pills. The YOLOv5-tiny model demonstrated the best detection accuracy, and efficientNet_b0 achieved the highest classification performance. While the model showed strong results, its generalization is limited by sample size and drug variability. Nonetheless, this approach holds promise for enhancing medication safety and reducing errors in outpatient care.
คณะเทคโนโลยีการเกษตร
Dwarf whipray (Brevitrygon heterura) is a common species found in a local market in the Gulf of Thailand. However, like many other species of stingrays, it is threatened by overfishing and habitat destruction. Therefore, an accurate species identification is crucial because conservation efforts may vary depending on the species. This study aims to understand morphological variation of B. heterura in the Gulf of Thailand by morphometric study and genetic analysis. During October 2022 and February 2023, we obtained 49 samples from research vessels fish landing ports and local fish markets. We observed two distinct groups based on 43 morphological variables/ratios. B. heterura samples from Chanthaburi, Rayong, Chonburi, Samut Sakhon, Nakhon Si Thammarat and Songkla provinces, called “group A," typically have longer snout length than those from Prachuap Khiri Khan provinces, called “group B" according to external morphological characters for species identification. Three morphological variables/ratios were significantly different between groups A and B. Main characters to explain intraspecific variations between group A and group B are further discussed. DNA barcoding based on a fragment of the cytochrome c oxidase subunit I (COI) gene were obtain from eight samples of group A and eight samples from group B. Pairwise percent sequence divergence (p-distance) for COI between group A and group B were 0.0-2.5. This study contributes to the understanding of variations of morphology and genetics of B. heterura in the Gulf of Thailand.