dc.contributor.author |
Aich, Sourav |
|
dc.date.accessioned |
2022-01-24T07:17:44Z |
|
dc.date.available |
2022-01-24T07:17:44Z |
|
dc.date.issued |
2019-07 |
|
dc.identifier.citation |
38p. |
en_US |
dc.identifier.uri |
http://hdl.handle.net/10263/7248 |
|
dc.description |
Dissertation under the supervision of Dr. Sushmita Mitra |
en_US |
dc.description.abstract |
Diabetic Retinopathy is the leading cause of blindness in today's modern world. Early
detection of Diabetic Retinopathy is crucial for it's prevention. To speed up the pro-
cess of detection, automated systems needs to be developed, which can grade a Fundus
Image for DR, without any human intervention. In this project we have used an
advanced variant of CNN(Convolutional Neural Network) integrated with Visual At-
tention Mechanism, for grading the Fundus Image for DR. We have also detected the
lesions such as Microaneurysms, Haemmorhage, Hard Exudates and Soft Exudates
in the Fundus images, and delineated their boundaries in the Fundus image. Finally
we have developed a joint segmentation and classi cation pipeline, which mimics a
pathologists action while grading a Fundus image. The system detects all the patholo-
gies in the Fundus Images, marks them and with this pathological information grades
the image. |
en_US |
dc.language.iso |
en |
en_US |
dc.publisher |
Indian Statistical Institute, Kolkata |
en_US |
dc.relation.ispartofseries |
Dissertation;;2019:2 |
|
dc.subject |
convolutional neural network (CNN or ConvNet) |
en_US |
dc.title |
Deep Attention Model for Diabetic Retinopathy Grading |
en_US |
dc.type |
Other |
en_US |