แจ้งเอกสารไม่ครบถ้วน, ไม่ตรงกับชื่อเรื่อง หรือมีข้อผิดพลาดเกี่ยวกับเอกสาร ติดต่อที่นี่ ==>
หากไม่มีอีเมลผู้รับให้กรอก thailis-noc@uni.net.th
หากไม่มีอีเมลผู้รับให้กรอก thailis-noc@uni.net.th
Sukawin Nunan. Image segmentation and classification of abnormal red blood cell based on morphology. Master's Degree(Data Science and Analytics). King Mongkut's Institute of Technology Ladkrabang. Central Library. : King Mongkut's Institute of Technology Ladkrabang, 2024.
Image segmentation and classification of abnormal red blood cell based on morphology
การแบ่งภาพและจำแนกความผิดปกติของเซลล์เม็ดเลือดแดงด้วยสัณฐานวิทยา
Name: Sukawin Nunan
LCSH:
Erythrocytes
LCSH:
Blood cellsBlood cells
-- Physiology
Abstract:
Abstract: Blood is a fluid that circulates throughout the body, accounting for approximately 7-8% of body weight. It is responsible for transporting oxygen, nutrients, and water to nourish tissues throughout the body, as well as removing waste products from various parts of the body. เท hematology, the study of red blood cell morphology has revealed that abnormalities in the shape and color of red blood cells are associated with potential diseases. This study developed three convolutional neural network models, ResNet50, lnceptionV3, and VGG16 architectures, as classifiers for red blood cell abnormality classification. The models were trained using 13 categories of red blood cell images obtained from image segmentation using the Watershed Segmentation technique. The dataset was divided into three types: data without data imbalance handling, data handling with weighted method, and data handling with SMOTE method. The results showed that the convolutional neural network model using the ResNet50 architecture trained on SMOTE-processed data achieved the best performance in classifying red blood cell abnormalities from shape and color, with an accuracy of 91.44%. Fine-tuning this model a dropout rate of 0.20 further improved the accuracy to 95.59%.
King Mongkut's Institute of Technology Ladkrabang. Central Library
Address:
BANGKOK
Email:
Lifelong@kmitl.ac.th
Name: Jiraphat Yokrattanasak
Role:
Thesis Advisor
Email :
jiraphat.yo@kmitl.ac.th
Created:
2024
Modified:
2024-11-19
Issued:
2024-11-19
วิทยานิพนธ์/Thesis
application/pdf
eng
DegreeName: Master of Science
Level: Master's Degree
Descipline: Data Science and Analytics
©copyrights King Mongkut's Institute of Technology Ladkrabang
RightsAccess:
ใช้เวลา
0.027326 วินาที
Sukawin Nunan
| Title | Contributor | Type |
|---|---|---|
| Image segmentation and classification of abnormal red blood cell based on morphology
สถาบันเทคโนโลยีพระจอมเกล้าเจ้าคุณทหารลาดกระบัง Sukawin Nunan | Jiraphat Yokrattanasak | วิทยานิพนธ์/Thesis |
Jiraphat Yokrattanasak