Journal
Automated Diagnosis of Pigmented Skin Lesions Based on Texture Characteristics
| Identifier | --bgohilavan-536 |
|---|---|
| Penulis / Pengarang | B.Gohila Vani, A.Selvaraj |
| Kata Kunci | Texture distinctiveness skin lesion segmentation Feature selection Probabilistic Neural Network classification. |
| Tanggal Publikasi | 08 Mar 2015 |
| Bahasa | Bahasa Indonesia |
| Abstrak / Deskripsi | Skin cancer is the most common type of cancer and represents 50% all new cancer detected every year. Melanoma is the most dangerous, deadly form of Skin Cancer. About 65% of Melanoma is attributed to UV damage done by exposure to the sun, much of which may not be easily apparent. The dermatologists cost to observe every patient. There is a necessity for automated diagnosis to calculate a patient’s risk of melanoma captured by standard digital camera. To distinguishing between malignant and benign tumor is a difficult even for proficient specialists. In Proposed method, computer-aided diagnosis system is useful tool to detect tumor. Usually, the process leads by pre-processing the skin lesion image, i.e. eliminating undesired artifacts such as freckles, hair or shadowing effects. Next, it executes a segmentation step to detect the lesion boundaries. Finally, lesion is identified based on the image part, several features are extracted and a classification is provided. We compare the segmentation outcomes to categorize which techniques have more exact results, and discuss how these outcomes may guidance in the following stages: the feature extraction and the classification of skin lesions. It achieves the total texture distinctiveness skin lesion segmentation accuracy (98.7%) and probabilistic neural network classification of skin lesions is exactness. |
File Lampiran / Download (1)
| No | Nama Berkas | Format | Ukuran | Aksi |
|---|---|---|---|---|
| 1 |
VOL3I2_938-945.pdf
Skin cancer is the most common type of cancer and represents 50% all new cancer detected every year. Melanoma…
|
application/pdf | 588,1 KB | ⬇️ Buka / Download |
Komentar & Diskusi (0)
Belum ada komentar untuk koleksi ini. Jadilah yang pertama berdiskusi!