readme

Train Validation set ( each has 7 corneal maps)

150 NOR
150 KCN
123 Suspect
Test set ( each has 7 corneal maps)

50 NOR
50 KCN
50 Suspect
Datasets and pre-processing
The protocol of the study (0094/2020) was approved by the Institutional Review Board of Federal University of São Paulo - UNIFESP/EPM as coordinator center and Hospital de Olhos-CRO, Guarulhos, as side center. Corresponding data use agreements were signed among parties to use the data. The study was conducted in accordance with ethical standards in the declaration of Helsinki and its later amendments. If required, respective informed consent was obtained from participants and the data was de-identified in Brazil before any further processing.
Three corneal specialists (including RMH) conducted vision tests and ophthalmic examinations under standard conditions and collected corneal images using Scheimpflug imaging systems (Pentacam, Oculus Optikgera¨te GmbH). There were three corneal trained specialists who performed eye classification. We dealt with disagreements favoring two versus one vote. The clinicians were instructed to grade each eye as normal, suspected KCN, or KCN. Eyes were labeled as a KCN suspect based on standard criteria in earlier studies. More specifically, eyes were labeled as suspected KCN if corneal topography included atypical, localized steepening or an asymmetrical bowtie pattern. Eyes were labeled as suspected KCN if the keratometric curvature was greater than 47.00 D, oblique cylinder more than 1.50 D or central corneal thickness below 500 microns. Each eye of the patient was evaluated independently. Furthermore, raw data on the elevation maps, including Belin –Ambrosio Ectasia Display (BAD-D) indices, Progression Thickness Increase (PTI) represented by corneal thickness spatial profile (CTSP) and percentage of PTI. The Belin ABCD progression display was also examined. Eyes were labeled as suspected KCN if there was abnormal front elevation, high PTI, or abnormal BAD-D.
The development (training) dataset included corneal images collected using different Pentacam instruments with different settings (different color scale steps of the maps compared to the previous subset). All color scales were based on decimal scale grading using microns for corneal thickness, and elevation maps and diopters for axial/sagittal curvature maps. Additional independent dataset, collected from a different clinic in Brazil, was also used to validate the proposed hybrid DL approach.
A total of 204 eyes of 104 patients were normal (the group was represented as NOR), 215 eyes of 113 patients had KCN, and 123 eyes of 63 patients were suspected KCN (SUSPECT). The mean age (± SD) of the subjects in the normal, KCN, and suspected KCN were 33.4 (±10.1), 29.0 (±9.3), and 28.6 (±9.4) years, respectively. Images from 56 normal eyes and 58 eyes with KCN were collected from a Pentacam instrument with settings different from others.
The independent validation subset included 150 eyes of 85 patients collected from de Olhos-CRO private hospital (Guarulhos, SP, Brazil). This dataset included 50 normal eyes from 29 subjects, 50 KCN eyes from 31 patients, and 50 suspect KCN eyes from 25 patients. The mean age (± SD) of the subjects in the normal, KCN, and suspected KCN were 29.5 (±4.7), 26.3 (±6.8), and 29.1 (±5.3) years, respectively.


View less

Image

Computer Vision

Eyes and Vision

Image Classification
Usability
7.50

License
Other (specified in description)


Expected update frequency
Never

Independent Test Set(1 directories)
About this directory
This file containts test images which is separed in 3 labels 'keratoconus' , 'normal' , 'suspect' which each corneal have 7 maps

Test set ( each has 7 corneal maps)

50 NOR

Source: https://www.kaggle.com/datasets/elmehdi12/keratoconus-detection