What is it about?
Microaneurysm (MA) is the earliest lesion of diabetic retionpathy. This paper proposes a deconvolutional neural network to accurately discriminate MA from non-MA. Experimental results demonstrate that, the proposed method is able to achieve significant sensitivity and accuracy on multiple public datasets, in comparison to the state-of-the-art.
Featured Image
Photo by v2osk on Unsplash
Why is it important?
An efficient network is proposed to accurately detect MA in retinal funds image. Compared to the traditional CNN using a large number of training samples, the random under-sampling based SODNet can detect small object with imbalanced data, in terms of few training samples.
Perspectives
We hope that this article can not only detect MA in retinal image, but also detect other small objects in various kinds of images, containing few training samples.
Xinpeng Zhang
Guangdong University of Technology
Read the Original
This page is a summary of: SODNet: Small Object Detection Using Deconvolutional Neural Network, IET Image Processing, April 2020, the Institution of Engineering and Technology (the IET),
DOI: 10.1049/iet-ipr.2019.0833.
You can read the full text:
Contributors
The following have contributed to this page
