Scholarlink

The amount of people with diabetes is rising quickly in India. Numerous illnesses, including neuropathy, renal failure, heart attacks, eyesight loss, and skin conditions are becoming more common among diabetic people. We are currently concentrating on visual loss in this research. Prompt sickness identification and management can help identify visual loss. Vision loss is caused by diabetic retinopathy, and one of the disease’s early side effects is microaneurysms (MAs). Early on in the development of diabetic retinopathy, the fundus evaluation is highly beneficial. However, because MAs frequently appear as tiny faint patches, it might be difficult for medical professionals to recognize MAs on retinal images. As a result, a lot of research has been done on robotized MA recognition. The double-ring filter, the index of shape on the Hessian structure, and the Gabor filter are all techniques that are currently in use, and this study suggests a single MA identifier by combining them. However, this study uses deep convolutional neural networks (DCNN) to do automated MA identification because DCNN has demonstrated superior performance in photo acknowledgment tests. We structured the suggested approach using a three-layer insight with 48 elements for false positives (FPs) reduce features and a two-step DCNN. The major DCNN in the two-step DCNN is used to begin MA discovery, while the second DCNN is used to minimize FPs. Utilizing the DIARETDB1 data base, the suggested approach demonstrates flawless implementation. In order to guarantee the AI model’s accuracy and efficacy in identifying and diagnosing diabetic retinopathy, it can be continuously enhanced over time. Deep learning algorithms can be used to train the AI model to recognize the MAs.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Analysis on Deep Convolution Neural Network Evaluation of Microaneurysms Identification on Retinal

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • A. Srinivasula Reddy,
  • X. S. Asha Shiny,
  • Y. Aruna Suhasini Devi,
  • G. Karthik Reddy,
  • G. Srikanth,
  • B. P. Deepak Kumar

摘要

The amount of people with diabetes is rising quickly in India. Numerous illnesses, including neuropathy, renal failure, heart attacks, eyesight loss, and skin conditions are becoming more common among diabetic people. We are currently concentrating on visual loss in this research. Prompt sickness identification and management can help identify visual loss. Vision loss is caused by diabetic retinopathy, and one of the disease’s early side effects is microaneurysms (MAs). Early on in the development of diabetic retinopathy, the fundus evaluation is highly beneficial. However, because MAs frequently appear as tiny faint patches, it might be difficult for medical professionals to recognize MAs on retinal images. As a result, a lot of research has been done on robotized MA recognition. The double-ring filter, the index of shape on the Hessian structure, and the Gabor filter are all techniques that are currently in use, and this study suggests a single MA identifier by combining them. However, this study uses deep convolutional neural networks (DCNN) to do automated MA identification because DCNN has demonstrated superior performance in photo acknowledgment tests. We structured the suggested approach using a three-layer insight with 48 elements for false positives (FPs) reduce features and a two-step DCNN. The major DCNN in the two-step DCNN is used to begin MA discovery, while the second DCNN is used to minimize FPs. Utilizing the DIARETDB1 data base, the suggested approach demonstrates flawless implementation. In order to guarantee the AI model’s accuracy and efficacy in identifying and diagnosing diabetic retinopathy, it can be continuously enhanced over time. Deep learning algorithms can be used to train the AI model to recognize the MAs.