Glaucoma Detection through a Novel Hyperspectral Imaging Band Selection and Vision Transformer Integration

被引:2
|
作者
Wang, Ching-Yu [1 ]
Nguyen, Hong-Thai [2 ]
Fan, Wen-Shuang [1 ]
Lue, Jiann-Hwa [3 ]
Saenprasarn, Penchun [4 ]
Chen, Meei-Maan [5 ]
Huang, Shuan-Yu [3 ]
Lin, Fen-Chi [6 ]
Wang, Hsiang-Chen [3 ,7 ]
机构
[1] Dalin Tzu Chi Hosp, Buddhist Tzu Chi Med Fdn, Dept Ophthalmol, Chiayi 62247, Taiwan
[2] Natl Chung Cheng Univ, Dept Mech Engn, Chiayi 62102, Taiwan
[3] Cent Taiwan Univ Sci & Technol, Dept Optometry, 666 Buzih Rd,, Taichung 406053, Taiwan
[4] Shinawatra Univ, Sch Nursing, 99 Moo 10, Samkhok 12160, Pathum Thani, Thailand
[5] Natl Chung Cheng Univ, Ctr Innovat Res Aging Soc CIRAS, 168 Univ Rd, Chiayi 62102, Taiwan
[6] Kaohsiung Armed Forces Gen Hosp, Dept Ophthalmol, 2,Zhongzheng 1 Rd, Kaohsiung 80284, Taiwan
[7] Hitspectra Intelligent Technol Co Ltd, Kaohsiung 80661, Taiwan
关键词
glaucoma detection; hyperspectral imaging; vision transformer; NERVE-FIBER LAYER; OXYGEN-SATURATION; OPTIC DISC; DIAGNOSIS; PROGRESSION; PATTERN; IMAGES;
D O I
10.3390/diagnostics14121285
中图分类号
R5 [内科学];
学科分类号
1002 ; 100201 ;
摘要
Conventional diagnostic methods for glaucoma primarily rely on non-dynamic fundus images and often analyze features such as the optic cup-to-disc ratio and abnormalities in specific retinal locations like the macula and fovea. However, hyperspectral imaging techniques focus on detecting alterations in oxygen saturation within retinal vessels, offering a potentially more comprehensive approach to diagnosis. This study explores the diagnostic potential of hyperspectral imaging for glaucoma by introducing a novel hyperspectral imaging conversion technique. Digital fundus images are transformed into hyperspectral representations, allowing for a detailed analysis of spectral variations. Spectral regions exhibiting differences are identified through spectral analysis, and images are reconstructed from these specific regions. The Vision Transformer (ViT) algorithm is then employed for classification and comparison across selected spectral bands. Fundus images are used to identify differences in lesions, utilizing a dataset of 1291 images. This study evaluates the classification performance of models using various spectral bands, revealing that the 610-780 nm band outperforms others with an accuracy, precision, recall, F1-score, and AUC-ROC all approximately at 0.9007, indicating its superior effectiveness for the task. The RGB model also shows strong performance, while other bands exhibit lower recall and overall metrics. This research highlights the disparities between machine learning algorithms and traditional clinical approaches in fundus image analysis. The findings suggest that hyperspectral imaging, coupled with advanced computational techniques such as the ViT algorithm, could significantly enhance glaucoma diagnosis. This understanding offers insights into the potential transformation of glaucoma diagnostics through the integration of hyperspectral imaging and innovative computational methodologies.
引用
收藏
页数:16
相关论文
共 50 条
  • [1] HybridGT: An Integration of Graph Transformer and LSTM for Effective Hyperspectral Band Selection
    Neela, Nagaraju
    Veerakumar, T.
    Panda, Manoj Kumar
    Subudhi, Badri Narayan
    Esakkirajan, S.
    Bouwmans, Thierry
    INTERNATIONAL JOURNAL OF REMOTE SENSING, 2024,
  • [2] Hyperspectral Band Selection for Human Detection
    Uto, Kuniaki
    Kosugi, Yukio
    Murase, Toru
    Takagishi, Sigenori
    2012 IEEE 7TH SENSOR ARRAY AND MULTICHANNEL SIGNAL PROCESSING WORKSHOP (SAM), 2012, : 501 - 504
  • [3] Hyperspectral anomaly detection with vision transformer and adversarial refinement
    Xu, Yating
    Zhao, Kai
    Zhang, Liangang
    Zhu, Mengyao
    Zeng, Dan
    INTERNATIONAL JOURNAL OF REMOTE SENSING, 2023, 44 (13) : 4034 - 4057
  • [4] Underwater Hyperspectral Target Detection with Band Selection
    Fu, Xianping
    Shang, Xiaodi
    Sun, Xudong
    Yu, Haoyang
    Song, Meiping
    Chang, Chein-I
    REMOTE SENSING, 2020, 12 (07)
  • [5] Selection of optimal combinations of band-pass filters for ice detection by hyperspectral imaging
    Nakauchi, Shigeki
    Nishino, Ken
    Yamashita, Takuya
    OPTICS EXPRESS, 2012, 20 (02): : 986 - 1000
  • [6] BSFormer: Transformer-Based Reconstruction Network for Hyperspectral Band Selection
    Liu, Yufei
    Li, Xiaorun
    Xu, Zezhong
    Hua, Ziqiang
    IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, 2023, 20
  • [7] Band selection for oxygenation estimation with multispectral/hyperspectral imaging
    Ayala, Leonardo
    Isensee, Fabian
    Wirkert, Sebastian J.
    Vemuri, Anant S.
    Maier-Hein, Klaus H.
    Fei, Baowei
    Maier-Hein, Lena
    BIOMEDICAL OPTICS EXPRESS, 2022, 13 (03) : 1224 - 1242
  • [8] Comparison of unsupervised band selection methods for hyperspectral imaging
    Martinez-Uso, Adolfo
    Pla, Filiberto
    Sotoca, Jose M.
    Garcia-Sevilla, Pedro
    PATTERN RECOGNITION AND IMAGE ANALYSIS, PT 1, PROCEEDINGS, 2007, 4477 : 30 - +
  • [9] Detection of yellow starthistle through band selection and feature extraction from hyperspectral imagery
    Miao, Xin
    Gong, Peng
    Swope, Sarah
    Pu, Ruiliang
    Carruthers, Raymond
    Anderson, Gerald L.
    PHOTOGRAMMETRIC ENGINEERING AND REMOTE SENSING, 2007, 73 (09): : 1005 - 1015
  • [10] Multiple Band Selection for Anomaly Detection in Hyperspectral Imagery
    Wang, Lin
    Chang, Chein-I
    2016 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM (IGARSS), 2016, : 7022 - 7025