This document discusses an automated bleb appearance grading scale that uses stereo images and image processing algorithms. Glaucoma is the second leading cause of blindness worldwide and is treated through trabeculectomy surgery, which creates a new drainage channel for intraocular fluid. After surgery, a slightly raised bleb tissue forms where fluid drains. The Indiana Bleb Appearance Grading Scale is commonly used to assess bleb formation based on four parameters but relies on subjective manual assessment. The proposed method uses stereo images and extracts height bleb features from disparity pixels while classifying features of three parameters using K-NN algorithms. This increased the Cohen's kappa coefficient between observers from 0.21 to 0.36, indicating better agreement in
This document discusses an automated bleb appearance grading scale that uses stereo images and image processing algorithms. Glaucoma is the second leading cause of blindness worldwide and is treated through trabeculectomy surgery, which creates a new drainage channel for intraocular fluid. After surgery, a slightly raised bleb tissue forms where fluid drains. The Indiana Bleb Appearance Grading Scale is commonly used to assess bleb formation based on four parameters but relies on subjective manual assessment. The proposed method uses stereo images and extracts height bleb features from disparity pixels while classifying features of three parameters using K-NN algorithms. This increased the Cohen's kappa coefficient between observers from 0.21 to 0.36, indicating better agreement in
This document discusses an automated bleb appearance grading scale that uses stereo images and image processing algorithms. Glaucoma is the second leading cause of blindness worldwide and is treated through trabeculectomy surgery, which creates a new drainage channel for intraocular fluid. After surgery, a slightly raised bleb tissue forms where fluid drains. The Indiana Bleb Appearance Grading Scale is commonly used to assess bleb formation based on four parameters but relies on subjective manual assessment. The proposed method uses stereo images and extracts height bleb features from disparity pixels while classifying features of three parameters using K-NN algorithms. This increased the Cohen's kappa coefficient between observers from 0.21 to 0.36, indicating better agreement in
Agwin Fahmi Fahanani1, Hasballah Zakaria2, Andika Prahasta3, Elsa Gustianty4, R. Maula Rifada5, Astrid Chairini6 1,2 Department of Biomedical Engineering, Institut Teknologi Bandung, Bandung, Indonesia (E-mail: agwinfahmi@gmail.com1, hasballah@stei.itb.ac.id2) 3,4,5,6 Faculty of Medicine, Padjajaran University, Bandung, Indonesia (E-mail: andikaprahasta@yahoo.com3, elsagustianty2017@gmail.com4, maularmd@gmail.com5, atid_gtct@yahoo.com6) Abstract— According to World Health Organizaton (WHO), the second leading blindness caused in the world is glaucoma. Glaucoma is caused by builded up of intraocular fluid due to the blocked drainage channel. In consequently, intraocular pressure is increasing, so it presses the nerve cells. The most common treatment for glaucoma is trabeculectomy surgery, that provides a new drainage channel so the intraocular fluid can flow properly and reabsorb. After the surgery, a slightly lifted tissue due to flowing fluid, called bleb, will appear. The examination of bleb formation is necessary to calculate the succesful rate of the surgery. Indiana Bleb Appereance Grading Scale (IBAGS) is the most common standard to asses the bleb formation. This standard provides four parameters to examine, bleb height, bleb extent, bleb vascularity, and leakage. Ophthalomologist uses this standard to grade patients manually, so the result is very subjective. This work proposed an image processing algorithm using stereo images database to produce better depth perception for the assesment. The height bleb feature was extracted from disparity pixels, while feature from other parameters were calculated afterwards. The K-NN algorithm used to classify the features from three parameters based on the IBAGS standard. The result showed Cohen’s kappa coefficient rate from three parameters was increased from 0.21 to 0.36 indicating a better agreement between inter- observers assesment. Keywords—glaucoma, trabeculectomy, bleb, Indiana Bleb Appereance Grading Scale (key words)