Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

A typical colored fundus image.

More »

Fig 1 Expand

Fig 2.

Research framework and workflow pipeline.

More »

Fig 2 Expand

Fig 3.

Class label distribution for the study.

More »

Fig 3 Expand

Table 1.

Image augmentation parameters.

More »

Table 1 Expand

Fig 4.

DIFE from the backbone architecture.

More »

Fig 4 Expand

Fig 5.

PFMR module.

More »

Fig 5 Expand

Fig 6.

Architecture of the LCC module.

More »

Fig 6 Expand

Table 2.

Layer-wise specification of the LiteFeatNet model.

More »

Table 2 Expand

Table 3.

Comparison of number and size of parameters.

More »

Table 3 Expand

Table 4.

Target deployment platforms.

More »

Table 4 Expand

Fig 7.

Accuracy plots.

(A) CNN-1 [17] (B) DenseNet121 (C) DeB5-XNet [24] (D) DIA-VXNET [29] (E) LightCNN [16] (F) MobileNet (G) MobileNetV2 (H) NAS_m1 [30] (I) NAS_m2 [30] (J) NASNetMobile (K) Proposed Model (LiteFeatNet) (L) Swin V2-Small (M) Swin V2-Tiny.

More »

Fig 7 Expand

Fig 8.

Loss plots.

(A) CNN-1 [17] (B) DenseNet121 (C) DeB5-XNet [24] (D) DIA-VXNET [29] (E) LightCNN [16] (F) MobileNet (G) MobileNetV2 (H) NAS_m1 [30] (I) NAS_m2 [30] (J) NASNetMobile (K) Proposed Model (LiteFeatNet) (L) Swin V2-Small (M) Swin V2-Tiny.

More »

Fig 8 Expand

Fig 9.

Confusion matrix for LiteFeatNet using RFMiD.

More »

Fig 9 Expand

Table 5.

Overall performance metrics for RFMiD.

More »

Table 5 Expand

Table 6.

Class-wise performance metrics for RFMiD.

More »

Table 6 Expand

Fig 10.

Class label distribution for RFMiD 2.0.

More »

Fig 10 Expand

Fig 11.

Pre-processing pipeline for RFMiD 2.0.

More »

Fig 11 Expand

Fig 12.

Evaluation mechanism for RFMiD 2.0.

More »

Fig 12 Expand

Fig 13.

Confusion matrix for LiteFeatNet using RFMiD 2.0.

More »

Fig 13 Expand

Table 7.

Evaluation metrics for RFMiD 2.0.

More »

Table 7 Expand

Fig 14.

External Dataset for Scalability Assessment.

More »

Fig 14 Expand

Table 8.

Class labels and the number of images supplemented for scalability assessment.

More »

Table 8 Expand

Table 9.

Performance metrics for 4 retinal pathologies.

More »

Table 9 Expand

Table 10.

Performance metrics for 5 retinal pathologies.

More »

Table 10 Expand

Table 11.

Ablation study experiments for the proposed LiteFeatNet.

More »

Table 11 Expand

Fig 15.

Performance evaluation of ablation experiments.

More »

Fig 15 Expand

Table 12.

Ablation experiments for other backbone architectures.

More »

Table 12 Expand

Fig 16.

Performance improvement with NASNetMobile backbone.

More »

Fig 16 Expand

Fig 17.

Performance improvement with DenseNet121 backbone.

More »

Fig 17 Expand

Fig 18.

Performance improvement with MobileNetV2 backbone.

More »

Fig 18 Expand

Fig 19.

Comparison of inference times.

More »

Fig 19 Expand