├── CMPB_OCT_Revised.pdf ├── CRF_OCT_SEG v1.zip └── README.md /CMPB_OCT_Revised.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/arunava555/OCT-layer-segmentation/f53f9ec3d4346967bd6612bd16798b33c9da67eb/CMPB_OCT_Revised.pdf -------------------------------------------------------------------------------- /CRF_OCT_SEG v1.zip: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/arunava555/OCT-layer-segmentation/f53f9ec3d4346967bd6612bd16798b33c9da67eb/CRF_OCT_SEG v1.zip -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # OCT-layer-segmentation: Segmentation of intra-retinal tissue layers in OCT B-scans 2 | Author: Arunava Chakravarty 3 | 4 | Supervisor: Prof. Jayanthi Sivaswamy 5 | 6 | Center for Visual Information Technology, IIIT-Hyderabad (https://cvit.iiit.ac.in/projects/mip/) 7 | 8 | The code is based on our following publications: 9 | 10 | [1] Arunava Chakravarty and Jayanthi Sivaswamy, "A Supervised Joint Multi-layer Segmentation Framework for Retinal Optical Coherence Tomography Images using Conditional Random Field", Computer Methods and programs in Biomedicine, 2018. https://www.sciencedirect.com/science/article/pii/S0169260717314645 11 | 12 | [2] Chakravarty, Arunava, and Jayanthi Sivaswamy. "End-to-End Learning of a Conditional Random Field for Intra-retinal Layer Segmentation in Optical Coherence Tomography." Medical Image Understanding and Analysis, pp. 3-14. Springer, Cham, 2017. https://link.springer.com/chapter/10.1007/978-3-319-60964-5_1 13 | 14 | 15 | We have developed a Conditional Random Field (CRF) based Joint Multi-layer segmentation of the intra-retinal layers in retinal OCT B-scans. The CRF is learned in an end-to-end manner using Structured Prediction. The test code along with the trained models for the experiments in [1] is being released publicly to allow other researchers to develop, compare and benchmark their algorithms. This code is not clinically approved and released for non-commerical research purposes only. Further details are available in the Readme.txt file within the downloadable zip file. 16 | 17 | Please note that this code is a slightly older version and uses different size of convolution filters as reported in [1]. 18 | However, the there is no significant difference in performance. 19 | --------------------------------------------------------------------------------