Hasil Pencarian  ::  Simpan CSV :: Kembali

"Computed Tomography (CT) merupakan sebuah pengembangan dari modalitas sinar-X dapat menghasilkan citra yang lebih jelas dan efektif untuk memberikan informasi abnormal pada organ, salah satu nya adanya nodul pada paru-paru. Nodul paru merupakan pertumbuhan jaringan abnormal pada paru yang digunakan sebagai diagnosis dini kanker paru. Umumnya, deteksi pertama nodul paru diperoleh dari citra CT yang didiagnosis secara visual oleh ahli radiologi. Artinya subjektivitas individu radiologis berpengaruh dalam citra diagnosis tersebut. Untuk membantu ahli radiologi dalam mendeteksi dan mengevaluasi nodul paru pada citra CT secara otomatis, penelitian ini telah mengembangkan sistem Computer-Aided Detection (CAD). Sistem CAD menggunakan metode segmentasi Otsu, dengan ekstraksi fitur Gray Level Co-occurrence Matrix (GLCM) sebagai input untuk klasifikasi nodul. Algoritma Random Forest digunakan untuk membedakan antara normal dan abnormal pada citra CT, khususnya citra dengan kelainan nodul paru. Evaluasi estimasi keberadaan nodul paru pada sistem menggunakan parameter-parameter Receiver Operating Characteristic (ROC) dengan menggunakan 207 citra pelatihan online, 96 citra pengujian online dan 146 citra pengujian lokal. Pada pengujian online diperoleh akurasi 92,7%, sensitivitas 95%, dan (AUC) 0,919. Untuk pengujian lokal diperoleh akurasi 89,7%, sensitivitas 94%, dan (AUC) 0,891. Hasil evaluasi menunjukkan bahwa sistem CAD yang dikembangkan baik digunakan untuk mengenali citra paru normal dan abnormal.

---

Computed Tomography (CT) is a development of X-ray modalities that can produce clearer and more effective images to provide abnormal information on organs, one of which is the presence of nodules in the lungs. Lung nodules are abnormal tissue growths in the lungs that are used as an early diagnosis of lung cancer. Generally, the first detection of lung nodule is obtained from CT images that visually diagnosed by radiologist. That means individual radiologist subjectivities influence in that image diagnoses. Reducing individual subjectivities, this work has developed computerized aided detection (CAD) system for evaluating lung nodules in CT images. The CAD system uses the Otsu segmentation method, with feature extraction of Gray Level Co-occurrence Matrix (GLCM) as input for nodule classification. Random forest algorithm is used to distinguish between normal and abnormal at CT images, particularly images with lung nodule abnormalities. Evaluation of the estimated presence of lung nodules in the system uses Receiver Operating Characteristic (ROC) parameters using 207 online training images, 96 online test images and 146 local test images. In online testing, the accuracy is 92.7%, sensitivity is 95%, and (AUC) is 0.919. For local testing, the accuracy is 89.7%, sensitivity is 94%, and (AUC) is 0.891. The evaluation results show that the developed CAD system is used to recognize normal and abnormal lung images."

"A state-of-the-art review of key topics in medical image perception science and practice, including associated techniques, illustrations and examples. This second edition contains extensive updates and substantial new content. Written by key figures in the field, it covers a wide range of topics including signal detection, image interpretation and advanced image analysis (e.g. deep learning) techniques for interpretive and computational perception. It provides an overview of the key techniques of medical image perception and observer performance research, and includes examples and applications across clinical disciplines including radiology, pathology and oncology. A final chapter discusses the future prospects of medical image perception and assesses upcoming challenges and possibilities, enabling readers to identify new areas for research. Written for both newcomers to the field and experienced researchers and clinicians, this book provides a comprehensive reference for those interested in medical image perception as means to advance knowledge and improve human health."

"The 33 revised full papers presented were carefully reviewed and selected from 67 submissions. The main aim of this workshop is to help advance the scientific research within the broad field of machine learning in medical imaging. It focuses on major trends and challenges in this area, and it presents work aimed to identify new cutting-edge techniques and their use in medical imaging."

"A virtopsy (virtual autopsy) is a minimally invasive alternative that can produce an efficient autopsy. The range of technologies employed in virtopsies include computer tomography, magnetic resonance imaging and spectroscopy, and 3D photogrammetry and surface scanning.Charred, badly decomposed, or mummified corpses, as well as those restrictions forced upon coroners by certain religious sects, often make autopsies impossible to perform. In addition, lack of manpower among the personnel charged with performing autopsies frequently creates a backlog of cases in the coroner's office. This delay increases the likelihood that causes of death will go undetermined and criminal perpetrators will go unpunished. The solution can be found in what has come to be known as the virtopsy®, a minimally invasive and efficient way to perform an autopsy."