CARNet: Cascade attentive RefineNet for multi-lesion segmentation of diabetic retinopathy images

Guo, Yanfei; Peng, Yanjun

doi:10.1007/s40747-021-00630-4

Cited by 28 publications

(17 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To confirm the effectiveness of the proposed method, we conduct a comparison between EDBNet and state-of-the-art on the AUPR metric (same as the one used in the IDRiD competition) as shown in Table 2. The comparison includes the top 5 teams on the IDRiD competition [16] (the first five rows in the table), as well as CARNet [4], EAD-Net [5], L-Seg [17], and SAA [7]. EDBNet surpasses all the state-of-the-art results of segmenting the SE retinal lesion segmentation by 1.55% of the AUPR metric.…”

Section: Comparison With Existing Methodsmentioning

confidence: 99%

“…We use a ResNet50 [10] encoder that was pre-trained on the ImageNet dataset as a backbone for our model. We selected this model because ResNet is the state-of-the-art backbone for many computer vision tasks [4,11].…”

Section: Backbone: the Encoder Layermentioning

confidence: 99%

“…Most of them use Convolutional Neural Networks (CNNs) like UNet [3] to automatically learn representative and high-level features from the input fundus images to achieve accurate segmentation. For instance, The authors of [4] proposed CARNet for multi-lesion segmentation. CARNet feeds the whole image and patch image into ResNet50 and ResNet101 networks, respectively, using a single attention refinement decoder.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

EDBNet: Efficient Dual-Decoder Boosted Network for Eye Retinal Exudates Segmentation

Ali

Abdel-Nasser

Valls

et al. 2022

Frontiers in Artificial Intelligence and Applications

View full text Add to dashboard Cite

Diabetic retinopathy (DR) is one of the most common causes of vision loss or blindness globally. Early detection of retinal eye lesions like hard exudates, soft exudates, microaneurysms, and hemorrhages is crucial to detect DR in a human eye. Therefore, accurate segmentation of lesions from eye fundus images is essential to develop efficient automated DR detection systems. This paper presented a novel hard and soft exudates lesions segmentation method called Efficient Dual-Decoder Boosted Network (EDBNet). EDBNet is composed of the following main components: 1) pre-trained ImageNet ResNet50 encoder with Atrous Spatial Pyramid Pooling (ASPP), 2) UNet decoder block with Gated Skip Connections mechanism to enhance capture more details of fundus images, 3) dual-decoder boosted to improve the performance segmentation of retinal lesion in the eye fundus images, and fusion outputs of the dual-decoder boosted to generate enhanced exudates segmentation. The effectiveness of the proposed framework is assessed on the IDRiD publicly dataset in terms of accuracy, Area Under Precision-Recall (AUPR), IOU, and Dice metrics. EDBNet obtains 99.8, 74.4, 78.0, and 87.6% of soft exudates, respectively. For hard exudates, EDBNet achieves 99.5, 85.3, 80.3, and 89.1%, respectively. The experimental results also demonstrate that EDBNet outperforms many state-of-the-art methods.

show abstract

Section: Comparison With Existing Methodsmentioning

confidence: 99%

Section: Backbone: the Encoder Layermentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

EDBNet: Efficient Dual-Decoder Boosted Network for Eye Retinal Exudates Segmentation

Ali

Abdel-Nasser

Valls

et al. 2022

Frontiers in Artificial Intelligence and Applications

View full text Add to dashboard Cite

show abstract

“…Yanfei Guo and Yanjun Peng [15] developed Cascade Attentive Refine Net (CARNet) for multi-lesion segmentation of DR images. The segmentation of lesion showed challenge due to the various sizes and complex structures.…”

Section: Literature Reviewmentioning

confidence: 99%

Enhanced Optimization using Advanced Cuckoo Search Algorithm with Ensemble classification for detection of Diabetic Retinopathy

Basavaraju¹,

Ganesarathinam²

2022

INDJCSE

View full text Add to dashboard Cite

Diabetic Retinopathy (DR) is an eye disease which occurs due to enormous glucose level in the blood among diabetic patients. The diabetic patients are having higher chances of getting blindness if the sugar level is increased in body. An identification of landmark features present in the fundus images has to accurately find the features from the optic disc of fundus images. The existing researches used various Artificial Intelligence (AI) techniques for screening and diagnosing the DR earlier to prevent the diabetic patients from blindness which was determined based on the level of DR severity. However, the existing models were efficient that consumed time and the premature convergence was resulted with the drawback for the real world optimization approach. Thus, the proposed Dynamic weighted based Cuckoo Search Algorithm (DWCSA) is modified to improve the performances based on their basic structure. The higher convergence helped to train the model better for finding solutions. The proposed model overcomes the constraint issues occurred during feature selection process showed improvement in the accuracy. The proposed DWCSA obtained the accuracy of 97.46 % compared to the existing CNN model that obtained 98.94 % of accuracy for DIARETDB1 whereas, the eophtha dataset obtained accuracy of 98.91% for the proposed model.

show abstract

“…Caixia et al 30 developed a model by combining U‐Net++ with a residual module that enables the network to segment the lesions more accurately by regaining the relevant low‐level features. Guo et al 31 implemented a cascade attentive RefineNet (CARNet) by combining U‐Nets. Whereas, the features of every decoder of one U‐Net are concatenated with the respective decoder of other U‐Net to attain accurate segment masks.…”

Section: Related Workmentioning

confidence: 99%

A deeply supervised maximum response texton basedSegNetfor simultaneous multi retinal lesion segmentation

Pappu¹,

Krishna²,

Biswal³

et al. 2022

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Diabetic Retinopathy (DR) is a diabetic mellitus complication that causes vision impairment and may lead to permanent blindness. The early signs of DR that appear on the retinal surface are microaneurysms, hemorrhages, hard exudates, and soft exudates. Hence the automatic detection of these retinal lesions assists in the early diagnosis of DR. This paper presents a novel deep learning model, MRT‐SegNet (Maximum Response Texton – Segmentation Network) for the automatic segmentation of different retinal lesions simultaneously along with the optic disc. In the proposed MRT‐SegNet, each encoder block consists of an MRT filter bank that extracts the textural feature maps of the retinal images and then fuses them with the local feature maps that are extracted from the traditional encoder block of the network. This fusion enables the network to segment the minute lesions from the retinal surface. The proposed model is evaluated on the IDRiD dataset and achieves a mean Area Under the Precision & Recall Curve (mAUC_PR) of 0.698 and AUC_PR scores of 0.495, 0.706, 0.823, 0.769 for microaneurysms, hemorrhages, hard exudates, and soft exudates respectively. The experimental results demonstrate that the MRT‐SegNet outperformed other multi retinal lesion segmentation models by achieving superior performance.

show abstract

CARNet: Cascade attentive RefineNet for multi-lesion segmentation of diabetic retinopathy images

Cited by 28 publications

References 40 publications

EDBNet: Efficient Dual-Decoder Boosted Network for Eye Retinal Exudates Segmentation

EDBNet: Efficient Dual-Decoder Boosted Network for Eye Retinal Exudates Segmentation

Enhanced Optimization using Advanced Cuckoo Search Algorithm with Ensemble classification for detection of Diabetic Retinopathy

A deeply supervised maximum response texton basedSegNetfor simultaneous multi retinal lesion segmentation

Contact Info

Product

Resources

About