biomimetics

Article
Lightweight Optic Disc and Optic Cup Segmentation Based on
MobileNetv3 Convolutional Neural Network
Yuanqiong Chen 1,2 , Zhijie Liu 2 , Yujia Meng 2 and Jianfeng Li 2, *

1 School of Computer Science and Engineering, Central South University, Changsha 410000, China;
yqchen@vip.163.com
2 School of Computer Science and Engineering, Jishou University, Zhangjiajie 427000, China
* Correspondence: ljf@jsu.edu.cn; Tel.: +86-13574318900

Abstract: Glaucoma represents a significant global contributor to blindness. Accurately segmenting

the optic disc (OD) and optic cup (OC) to obtain precise CDR is essential for effective screening.
However, existing convolutional neural network (CNN)-based segmentation techniques are often
limited by high computational demands and long inference times. This paper proposes an efficient
end-to-end method for OD and OC segmentation, utilizing the lightweight MobileNetv3 network as
the core feature-extraction module. Our approach combines boundary branches with adversarial
learning, to achieve multi-label segmentation of the OD and OC. We validated our proposed approach
across three public available datasets: Drishti-GS, RIM-ONE-r3, and REFUGE. The outcomes reveal
that the Dice coefficients for the segmentation of OD and OC within these datasets are 0.974/0.900,
0.966/0.875, and 0.962/0.880, respectively. Additionally, our method substantially lowers computa-
tional complexity and inference time, thereby enabling efficient and precise segmentation of the optic
disc and optic cup.

Keywords: glaucoma screening; optic disc and optic cup segmentation; convolutional neural network;
adversarial generative network

1. Introduction
Citation: Chen, Y.; Liu, Z.; Meng, Y.;
Li, J. Lightweight Optic Disc and Glaucoma is a chronic eye condition and ranks as the second-most prevalent cause of
Optic Cup Segmentation Based on blindness globally. Experts predict that glaucoma patients worldwide will increase from
MobileNetv3 Convolutional Neural 64.3 million in 2013 to 118 million in 2040 [1,2]. At present, there are three main methods
Network. Biomimetics 2024, 9, 637. for glaucoma diagnosis based on information technology: The first method involves mea-
https://doi.org/10.3390/ suring intraocular pressure (IOP), which tends to increase in glaucoma patients, due to
biomimetics9100637 an imbalance between intraocular fluid production and drainage. However, a significant
Academic Editor: Heming Jia
number of glaucoma patients exhibit minimal changes in IOP, leading to reduced diagnostic
accuracy using this approach. The second method employs visual field (VF) evaluation,
Received: 30 July 2024 which necessitates sophisticated medical equipment, and involves subjective diagnosis
Revised: 7 October 2024 steps. Although widely used, this method presents challenges, in terms of equipment
Accepted: 16 October 2024 requirements and subjective interpretation. The third method, image evaluation of the
Published: 18 October 2024
optic nerve head (ONH), predominantly relies on the analysis of digital fundus images
(DFI) for glaucoma diagnosis and is commonly employed in clinical practice. However,
these existing diagnostic methods are characterized by high costs and low efficiency, ren-
Copyright: © 2024 by the authors.
dering them unsuitable for large-scale glaucoma screening and diagnosis. Due to the
Licensee MDPI, Basel, Switzerland. swift advancement of information technology, auxiliary diagnostic technology is crucial for
This article is an open access article large-scale glaucoma diagnosis and screening. Excellent auxiliary diagnosis methods can
distributed under the terms and significantly reduce the cost of diagnosis while improving the accuracy of clinical diagnosis.
conditions of the Creative Commons Among these auxiliary diagnostic techniques, the vertical cup-to-disc ratio (CDR) serves as
Attribution (CC BY) license (https:// a widely accepted reference standard. Accurate segmentation of the optic disc (OD) and
creativecommons.org/licenses/by/ optic cup (OC) is essential for precise CDR measurement and forms the basis for reliable
4.0/). auxiliary glaucoma diagnosis.

Biomimetics 2024, 9, 637. https://doi.org/10.3390/biomimetics9100637 https://www.mdpi.com/journal/biomimetics

Biomimetics 2024, 9, 637 2 of 13

Recently, deep learning techniques have shown significant progress in the segmen-
tation of medical images. Compared to traditional methods [3], deep learning-based
segmentation approaches generally outperform conventional methods, in terms of ac-
curacy and efficiency. Numerous researchers have made notable advancements in this
field. For instance, reference [4] introduced a general encoder–decoder network for seg-
menting the optic disc (OD) and optic cup (OC), which includes a multi-scale weight-
shared attention (MSA) module and a densely connected depthwise separable convolution
(DSC) module. Additionally, reference [5] introduced an unsupervised domain adaptation
framework—namely, input and output space unsupervised domain adaptation
(IOSUDA)—to address performance degradation in joint OD and OC segmentation. Fur-
thermore, reference [6] utilized a deep learning method based on M-Net, which applies
a polar coordinate transformation to convert fundus images into a polar coordinate sys-
tem. Subsequently, the transformed images are processed using M-Net, a multi-label deep
network that generates a probability map containing OD and OC regions. To address the
issue of automatic OD center localization and long training time, reference [7] proposed an
approach utilizing the fully convolutional network FC-DenseNet. However, this method
has its limitations. In reference [8], an improved fully convolutional network (FCN) was
utilized for preprocessing and simultaneous segmentation of the OD and OC. Reference [9]
presented a transferable attention U-Net (TAU) model for OD and OC segmentation across
different fundus image datasets. This model incorporates two discriminators and atten-
tion modules, to localize and extract invariant features across datasets. Reference [10]
introduced EDDense-Net, a segmentation network for estimating the cup-to-disc ratio
(CDR) for glaucoma diagnosis. Using dense blocks and grouped convolutional layers, it
captures spatial information and reduces complexity. Evaluated on two public datasets,
EDDense-Net outperformed existing methods in accuracy and efficiency, aiding ophthal-
mologists in diagnosing glaucoma. Reference [11] proposed a modified U-Net model for
detecting the edges of the optic cup and disc in fundus images of glaucoma patients. The ap-
proach utilized edge detection and dilation techniques to enhance segmentation. A novel
boundary-enhanced adaptive context network known as BEAC-Net was introduced in
reference [12], designed for accurate segmentation of both the optic disc and the optic
cup in retinal fundus images. BEAC-Net integrates an efficient boundary pixel attention
(EBPA) module and an adaptive context module (ACM), to improve boundary detection
and capture contextual information. Reference [13] presented LC-MANet, a multiplex
aggregation network for joint segmentation of the optic disc (OD) and optic cup (OC), to
estimate the cup-to-disc ratio (CDR) for glaucoma diagnosis. It integrates independent and
joint segmentation with a coarse-to-fine approach and uses multi-channel fusion to reduce
interference. The method outperforms existing techniques in accuracy on the RIM-ONE,
Drishti-GS, and REFUGE datasets.
Despite these advances, deep learning-based optic disc (OD) and optic cup (OC)
segmentation still faces several challenges. Firstly, the structure of the segmentation models
is usually complicated, which leads to high computation cost and long segmentation time.
Maintaining a balance between segmentation accuracy and computational cost and time
remains a challenge. Secondly, the model has limited generalization ability on different
datasets, making it difficult to achieve consistent performance on diverse datasets. Finally,
the model’s complexity and the substantial size of its weight files pose challenges for
deployment and use on mobile devices with limited resources.
In this paper, we present an end-to-end network model for optic disc (OD) and optic
cup (OC) segmentation that addresses existing challenges by minimizing computational
demands and accelerating inference speed, thus enabling efficient deployment on mobile
devices. Our methodology incorporates joint multi-label segmentation of the optic disc
(OD) and optic cup (OC), complemented by an additional boundary branch aimed at
improving segmentation accuracy. Furthermore, the use of adversarial learning techniques
allows us to refine segmentation boundaries, further boosting accuracy. When compared to
Biomimetics 2024, 9, 637 3 of 13

prior methods, our network model offers advantages including reduced parameter count,
lower computational overhead, quicker inference, and improved segmentation accuracy.
The principal contributions of this paper are outlined as follows: (1) We present a
comprehensive network for segmenting the optic disc and optic cup, designed to reduce
computational overhead and enhance inference speed. By incorporating a lightweight
feature-extraction network, we achieve improved segmentation efficiency while preserving
accuracy. (2) Our method ensures a swift inference process, with a processing time of
approximately 24 milliseconds, making it ideal for use in mobile devices and in clinical
settings. (3) By leveraging multi-label segmentation and adversarial learning techniques,
we refine boundary delineation, which boosts the segmentation accuracy for both the optic
disc and the optic cup and enhances the model’s overall generalization performance.

2. Methods
2.1. Generating the Network
The paper proposes a lightweight segmentation network that inherits the overall
structure of Deeplabv3+ [14]. The architecture of the MBG-Net network is illustrated in
Figure 1. After the feature-extraction network, it connects a spatial pyramid pooling module
(Atrous Spatial Pyramid Pooling, ASPP) with atrous convolution. Meanwhile, it integrates
more feature information by fusing the shallow features of the feature-extraction network
with the feature map acquired from the spatial pyramid pooling module, with the aim of
obtaining more comprehensive feature information. The difference is that the Deeplabv3+
network uses the ResNet [15] series as its feature-extraction network. ResNet networks have
excellent performance in feature extraction but come with a large amount of computation,
which is contrary to the goal of designing a lightweight and efficient segmentation network
model. Inspired by successful applications of the MobileNet [16–18] series on mobile
devices, this paper replaces ResNet with MobileNet as its feature-extraction network and
proposes a feature-extraction method based on MobileNetv3.

Figure 1. Overview of the MBG-Net network architecture: (a) is the feature-extraction module; (b) is
the boundary branch; (c) is the mask branch.

In terms of specific operations, MBG-Net utilizes the large version of the MobileNetv3
network as its feature-extraction network. However, only the first convolutional layer and
15 inverted residual blocks from the large MobileNetv3 version are utilized. Additionally,
the stride of the third-to-last inverted residual block is modified from 2 to 1. Experimental
comparisons have demonstrated that this alteration is advantageous for extracting local
features. To better extract contextual information, multi-scale feature fusion is performed
on the extracted feature maps, using the spatial pyramid pooling module with atrous
Biomimetics 2024, 9, 637 4 of 13

convolution. Subsequently, the feature map obtained from the spatial pyramid pooling
module undergoes operations like batch normalization (BN) and rectified linear unit
(ReLU) with a convolution kernel of size 1 × 1, to reduce the number of feature map
channels. Afterwards, the feature map is upsampled fourfold, and the resulting feature
map is combined with the feature map obtained from the third inverted residual block
of MobileNetv3.
To enhance the accuracy of OD and OC segmentation, a multi-label segmentation
method is used, where the resulting feature maps are input into the boundary prediction
branch and the mask prediction branch. Specifically, for the boundary branch, the fea-
ture map undergoes three convolution operations, where the output channels of the first
two convolutions are set to 256. The final convolution operation produces a feature map
with a single channel, resulting in the predicted boundaries for both the OD and OC.
For the mask branch, the feature map is concatenated with the predicted boundary feature
map and then subjected to a convolution operation. The output feature map consists of
two channels, which are in correspondence with the predicted masks for the optic disc
(OD) and optic cup (OC). The obtained mask features are then upsampled by a factor
of four, to generate a mask prediction map that matches the input image’s size. All the
aforementioned convolution operations use a stride of one. To enhance the accuracy of
boundary and mask predictions, the boundary regression loss and mask prediction loss are
formulated as defined in Equations (1) and (2), respectively:

N
1
Lb =
N ∑(yib − pib )2 (1)
i

N
1
Lm = −
N ∑[yim ∗ log( pim ) + (1 − yim ) log(1 − pim )] (2)
i

For Formula (1), N denotes the total number of pixels, yib is the natural boundary map
generated by morphological closure operation and Gaussian filter, and pib is the boundary
prediction map generated by the segmentation network. For Equation (2), yim is the ground-
truth mask label and pim is the mask prediction map predicted by the segmentation network.
The method in this paper sends the mask prediction map to the patch discriminator
(PatchGAN) to deceive the patch discriminator, in order to optimize the parameters of the
entire segmentation network, allowing the network to generate a more realistic prediction
map. The segmentation network is optimized using the subsequent adversarial loss:

N
1
L adv = −
N ∑ Lbce ( pm , 1) (3)
i

Joint Loss function: We can consider loss function that integrates both the segmentation
loss and the boundary loss as follows:

Ltotal = Lb + Lm + β ∗ L adv (4)

To summarize, the loss for the segmentation network comprises the boundary pre-
diction loss, the mask prediction loss, and the adversarial loss. β is the balance coefficient,
which is employed to balance the proportion of the adversarial loss function. Based on
empirical evidence, β is set to 0.01.

2.2. Adversarial Networks

A PatchGAN is employed as the adversarial network for MBG-Net. The concept of
adversarial learning aims to achieve optimal segmentation effects through a maximization
and minimization game between the generator and the discriminator. After comparing dis-
criminators, such as ImageGAN and PixelGAN, PatchGAN was chosen as the adversarial
generative network for MBG-Net. PatchGAN can capture local information in the output
Biomimetics 2024, 9, 637 5 of 13

space, allowing the segmentation network to emphasize local structural similarities within
image patches. This adversarial approach ensures that the segmentation masks adhere to
geometric constraints.
Concretely, a PatchGAN is connected after the mask branch. As illustrated in Figure 2,
the PatchGAN network encompasses five convolutional layers. The size of the convolu-
tional kernels is set to 4 × 4, with a stride of 2. The output channels of the five convolutional
layers increase progressively from the shallower to the deeper layers, with magnitudes
of 64, 128, 256, and 512, respectively. The final output channel is 2. The activation func-
tion following the last convolutional layer is the Sigmoid function, while for the other
convolutional layers, the activation functions are LeakyReLU with a negative slope value
of 0.2.

Figure 2. An overview of discriminator network architecture.

For the parameter training of the patch discriminator network, Equation (5) is used for
optimization, to distinguish whether the mask comes from the segmentation network or
the generator. An interactive training strategy between the generator and the discriminator
is employed during the training process, to optimize the parameters of the entire network.
A set of optimal model parameters is learned through the max–min game between the
generator and the discriminator.

N
1
LD = −
N ∑[ Lbce ( pm , 0) + Lbce (ym , 1)] (5)
i

In Equation (5), pm denotes the mask prediction map, while ym signifies the manually
annotated mask map.

3. Experiments and Results

3.1. Datasets
The experiment utilized three publicly accessible fundus datasets: Drishti-GS, RIM-
ONE-r3, and REFUGE. The Drishti-GS dataset consists of 101 fundus images, which include
31 normal and 70 glaucomatous images. The RIM-ONE-r3 dataset contains 159 fundus
images, with 85 normal and 74 glaucomatous images. As the largest open-source glaucoma
fundus dataset available to date, the REFUGE dataset contains 1200 fundus images divided
into three subsets: 400 for training, 400 for validation, and another 400 for testing purposes.
The sample graphs and annotation graphs of the three datasets are shown in Figure 3:
(a) and (d) show the glaucoma fundus sample map alongside its annotation from the Drishti-
GS dataset; (b) and (e) depict the normal eye fundus sample map and the corresponding
annotation map from the RIM-ONE-r3 dataset; while (c) and (f) depict both the glaucoma
fundus sample map and its annotation from REFUGE. In (d), (e), and (f), green indicates
the optic disc, whereas blue denotes the optic cup. The experiment used 320 images from
the REFUGE dataset for training and validation and 80 images for testing. The specific data
of the three datasets are shown in Table 1:
Biomimetics 2024, 9, 637 6 of 13

Table 1. Statistics of the datasets used in evaluating our method.

Dataset Number of Samples Image Size Cameras Release Year

Drishti-GS [19] 50 + 51 2047 × 1759 unknown 2014
RIM-ONE-r3 [20] 99 + 60 2144 × 1424 unknown 2015
REFUGE [21] 320 + 80 2124 × 2056 Zeiss Visucam 500 2018

Figure 3. The sample images and corresponding annotation images of the Drishti-GS, RIM-ONE-r3,
and REFUGE datasets are presented as follows: (a,d) represent the sample images and correspond-
ing annotations of the Drishti-GS dataset; (b,e) are the RIM-ONE-r3 dataset sample map and the
corresponding annotation map; (c,f) are the REFUGE dataset sample map and the corresponding
annotation map.

3.2. Implementation Details

The experimental model was built using the PyTorch deep learning framework. At first,
the segmentation network was trained, and, subsequently, an adversarial learning network
was introduced, to optimize the parameters of the entire MBG-Net network. A set of
optimal model parameters was obtained through alternate training. The Adam optimizer
was adopted for training the segmentation network, and the stochastic gradient descent
(SGD) algorithm was utilized when training the patch discriminator network. In the ex-
periment, the pre-trained weights on the ImageNet dataset were used as the initialization
weights of the feature-extraction network, namely, MobileNetv3_Large. The initial learning
rates of the segmentation network and the patch discriminator were set to 1 × 10−3 and
2.5 × 10−5 , respectively, and Momentum was set to 0.9. The experiments were trained
on an NVIDIA GTX 1080Ti GPU (NVIDIA Corporation, Santa Clara, CA, USA) with a
batch size of 8 for 300 epochs. For the region of interest (ROI) extraction, we followed
the method of [20]. First, we used the U-Net network for rough OD localization, then a
512 × 512 area centered on the OD mask was cropped, as input for MBG-Net. Due to the
limited availability of fundus images, which makes public datasets relatively small, we
employed several data augmentation techniques, to enhance the quantity and diversity of
the images. These techniques included random scaling, rotation, flipping, elastic transfor-
mation, contrast adjustment, noise addition, and random erasing. Additionally, to further
refine the segmentation results, we applied morphological operations and median filtering
Biomimetics 2024, 9, 637 7 of 13

to the prediction masks. This process involved techniques such as hole filling and selecting
the largest connected region, to achieve smoother and more natural boundaries.

3.3. Algorithm Evaluation

The Dice index (DI) is a standard evaluation index for the segmentation task, and the
CDR is one of the crucial indicators for clinical glaucoma screening. This paper adopted DI
and CDR to evaluate the segmentation performance of MBG-Net. The evaluation criteria
were defined as follows:

(2 × N_TP)
Dice = (6)
((2 × N_TP) + N_FP + N_FN )
VCD
CDR = (7)
VDD
N
δcdr = ∑ |CDRiS − CDRiG | (8)
i =1

N_TP, N_FP, and N_FN denoted the number of pixels corresponding to true pos-
itives, false positives, and false negatives, respectively. VCD and VDD represented the
vertical diameters of the cup and the disc, respectively, which were calculated from the
segmentation results of the cup and the disc. CDR represented the ratio of VCD and VDD.
CDRG and CDRS represented the vertical cup-to-disc ratio obtained from the true values
and the predicted segmentation, respectively. N represented the number of test samples.
The δcdr was defined as a measure of the precision of the CDR estimate, which calculated
the average error rate for all the samples, as shown in Equation (8). Lower values of δcdr
signified better prediction results.

3.4. Experimental Results

For the Drishti-GS and RIM-ONE-r3 datasets, we evaluated our MBG-Net against
several leading OD and OC segmentation methods: pOSAL [20], BEAL [22], TD-GAN [23],
BGA-net [24], Sevastopolsky [25], Zilly [26], BEAC-Net [12], LC-MANet [13], etc. The quan-
titative results are presented in Table 2, while the qualitative results are illustrated in
Figures 4 and 5.

Table 2. Performance comparison of attention blocks for optic disc and optic cup segmentation on
Drishti-GS and Rim_One_r3 datasets (mIoU was used to select the best model).

Datasets
Methods Drishti-GS Rim_One_r3
DIdisc DIcup δcdr DIdisc DIcup δcdr
U-net [19] 0.904 0.852 - 0.864 0.797 -
M_UNet [25] 0.95 0.85 - 0.95 0.82 -
M-Net [6] 0.967 0.808 - 0.952 0.802 -
CE-Net [27] 0.964 0.882 - 0.953 0.844 -
CDED-Net [28] 0.959 0.924 - 0.958 0.862 -
pOSAL [20] 0.965 0.858 0.082 0.865 0.787 0.081
Gan-based [23] 0.953 0.864 - 0.953 0.825 -
BEAL [22] 0.961 0.862 - 0.898 0.810 -
PDD-UNET [29] 0.963 0.848 0.105 0.970 0.876 0.066
BEAC-Net [12] 0.8614 0.8087 - 0.8582 0.7333 -
LC-MANet [13] 0.9723 0.9034 0.043 0.9729 0.8458 0.0444
Ours 0.974 0.900 0.045 0.966 0.875 0.043
Biomimetics 2024, 9, 637 8 of 13

Image
Ground Truth
Prediction

Figure 4. Prediction results of the Drishti-GS dataset (Image is the fundus picture, Ground Truth is
the annotation map, and Prediction is the network prediction map).
Image
Ground Truth
Prediction

Figure 5. Prediction results of the RIM-ONE-r3 dataset (Image is the fundus picture, Ground Truth is
the label map, and Prediction is the network prediction map).

For the REFUGE dataset, its training set was used as experimental data. We took
320 images as the training set and 80 as the test set. Compared to the BGA-net segmentation
model, which shows better performance in Table 3, the network model proposed in this
paper improved the inference time by about 16% and also enhanced the segmentation
accuracy of the OC. Additionally, the index also showed specific improvement. The quan-
titative results are shown in Table 3, and Figure 6 shows the qualitative results of the
REFUGE dataset.

Table 3. Performance comparison of Dice coefficient of OD and OC segmentation on REFUGE [30]

dataset.

Methods Ddisc Dcup δcdr Inference Time

U-net [19] 0.927 0.848 - -
pOSAL [20] 0.932 0.869 0.059 -
Psi-Net [21] 0.956 0.851 - -
BEAL [22] 0.945 0.860 - -
BGA-NET [24] 0.951 0.866 0.040 29.1 ms
ours 0.951 0.869 0.013 24.3 ms
Biomimetics 2024, 9, 637 9 of 13

Image
Ground Truth
Prediction

Figure 6. Prediction results of the REFUGE-train dataset (Image is the fundus picture, Ground Truth
is the annotation map, and Prediction is the network prediction map).

Additionally, this paper compared the model parameters, memory usage, compu-
tational cost, and inference time of the network models with different feature-extraction
networks, to demonstrate the excellent performance and lightweight nature of MobileNetv3
as a feature-extraction network. The specific results are shown in Table 4.

Table 4. Performance comparison of network models under different feature-extraction networks.

Backbone Total Params Total Memory Total MAdd Total Flops Total MemR + W Segmentation Time
efficientnetv2_m 54.9 M 2299.8 M 158.4 G 79.3 G 3.1 GB 192.7 ms
Xception 52.1 M 1524.5 M 165.8 G 83.0 G 3.1 GB 170.4 ms
Resnet50 38.4 M 845.3 M 138.5 G 69.3 G 1.7 GB 58.9 ms
Mobilenetv2 5.5 M 651.1 M 52.9 G 26.5 G 1.2 GB 29.1 ms
Ours 5.3 M 495.4 M 48.8 G 24.4 G 839.0 M 24.3 ms

The cup-to-disc ratio (CDR) is a crucial clinical parameter for diagnosing glaucoma,
and it serves as a fundamental basis for diagnosis by most ophthalmologists. In this paper,
the segmentation performance of BGA was evaluated by means of the receiver operating
characteristic (ROC) curve and the corresponding area under the curve (AUC). Generally
speaking, a higher AUC implies superior diagnostic performance, and greater accuracy
reflects the enhancement of algorithm performance. Figure 7 presents the ROC curves and
the corresponding AUC values on three public datasets, namely, Drishti-GS, RIM-ONE-r3,
and REFUGE-train.

Figure 7. ROC curve and AUC evaluation results of different datasets on BGA-Net.
Biomimetics 2024, 9, 637 10 of 13

3.5. Ablation Experiments

The subsequent ablation experiments aimed to assess the effectiveness of both the
boundary branch and the discriminator network: (1) baseline (feature-extraction network +
ASPP + mask branch); (2) baseline + boundary branch; (3) baseline + boundary branch +
discriminator network.
Ablation experiments were performed on the RIM-ONE-r3 dataset, and the experi-
mental results of the different module combinations are shown in Table 5. The experiments
demonstrated that, in comparison to the baseline network, both the boundary branch and
the discriminator network significantly enhanced the model’s segmentation performance,
to varying degrees. The network structure with boundary branches added to the baseline
improved the OC segmentation accuracy by 0.3% and decreased the mean absolute error
(δcdr ) by 0.004. Furthermore, when both the discriminator and the boundary branch compo-
nents were incorporated, the network structure of MBG-Net enhanced the OC accuracy by
0.7% compared to the baseline with boundary branch, and reduced the mean absolute error
(δcdr ) by 0.003, which proves the effectiveness of the boundary branch and the discriminator
network. Meanwhile, the qualitative segmentation results are given in Figure 8: green
indicates the optic disc (OD); blue represents the optic cup (OC); column (a) is the ROI area
corresponding to the fundus image; column (b) is the manual annotation map; column
(c) is the predicted segmentation map corresponding to the baseline; column (d) is the
predicted segmentation map corresponding to baseline + boundary branch; and column
(e) is the predicted segmentation map corresponding to baseline + boundary branch +
discriminator network (MBG-Net). Compared with columns (c) and (d), the boundary of
the segmentation map in column (e) is smoother and more natural, and it is closer to the
manual annotation map.
Table 5. Ablation paper on different components.

Dataset DIdisc DIcup δcdr

Baseline 0.964 0.865 0.050
Baseline + B 0.964 0.868 0.046
Baseline + B + G 0.966 0.875 0.043

(a) (b) (c) (d) (e)

Figure 8. Comparison of qualitative segmentation results under different module combinations:

(a) the ROI area of the fundus image; (b) the manual annotation map; (c) the predicted segmentation
map associated with the baseline model; (d) the predicted segmentation map for baseline + boundary
branch; and (e) the predicted segmentation map for baseline + boundary branch + discriminator
network (MBG-Net).
Biomimetics 2024, 9, 637 11 of 13

Furthermore, to verify the effectiveness of data preprocessing, we compared the

original images and the experimental images with ROI cropping on the Refuge dataset,
as shown in Table 6. The cropped dataset was able to effectively improve the OD and
OC Dice.

Table 6. The necessity of ROI region extraction on REFUGE [30] dataset.

Dataset Ddisc Dcup

Original dataset 0.920 0.820
ROI extraction dataset 0.962 0.880

4. Discussion
The CDR is an essential attribute in diagnosing glaucoma, and accurate segmentation
of the OD and OC is crucial to the precise acquisition of the CDR. In recent years, OD
and OC segmentation methods based on deep learning have made significant progress.
However, there is still a significant gap between research work and clinical application.
Most of the current segmentation networks have a large number of model parameters and
a long segmentation time, which cannot meet the clinical needs of mobile deployment and
real-time detection. This paper proposes a lightweight MBG-Net segmentation model for
the above problems. Our experiments show that the proposed method has reached an
advanced level, in terms of segmentation accuracy, computational cost, model parameters,
etc., indicating the application potential of the method in mobile deployment and real-
time detection.
Our experiments performed on the Drishti-GS, RIM-ONE-r3, and REFUGE datasets
demonstrated that the proposed method yielded the most advanced segmentation results
and the lowest absolute error δcdr with the least amount of parameters. Using the ROC
curve to evaluate the model’s performance, the accuracy rates on the three datasets were
91.90%, 84.16%, and 98.83%, respectively. In the experiment, the ROI region was extracted
from the original image according to the method in the literature, and the extracted ROI
region was used as the input of MBG-Net. To prove the necessity of ROI region extraction,
the following experiments were carried out on MBG-Net: keeping the original experimental
conditions unchanged, the original dataset REFUGE-train was trained and tested on MBG-
Net, and the results are shown in Table 3, which shows that the segmentation accuracy
of the OD decreased by 4.2% and that the segmentation accuracy of the OC decreased by
6%. This experiment shows the effectiveness and necessity of ROI extraction. In addition,
the inference time of this network for an image on a single NVIDIA GTX 1080Ti GPU was
only 24.3 ms, which shows the high efficiency and real-time performance of the MBG-
Net network.
To show the segmentation effect intuitively, this paper presents part of the segmenta-
tion renderings of the three datasets. The comparison shows that the segmentation effects
of this method on the three datasets are comparable to those of the manual annotation
images, which illustrates the generalization capability of this method.

5. Conclusions
This paper presents MBG-Net, an optimized network specifically developed for the
segmentation of the optic disc (OD) and optic cup (OC). MBG-Net specifically addresses
the challenges associated with OD and OC segmentation by integrating these tasks into a
unified framework. The network is optimized to deliver reduced training and computa-
tional expenses, quicker inference times, and enhanced segmentation accuracy. It leverages
MobileNetv3 for lightweight feature extraction and incorporates the boundary auxiliary
branches alongside adversarial learning techniques, to boost segmentation performance
while maintaining accuracy. We validated the effectiveness of MBG-Net through extensive
experiments on three widely used fundus datasets. Our results demonstrate that the net-
work not only achieves superior segmentation performance but also maintains low δcdr
Biomimetics 2024, 9, 637 12 of 13

values across all datasets, underscoring its efficacy in tackling OD and OC segmentation
challenges and supporting glaucoma diagnosis. In the future, we will produce a further
paper on how to deploy web applications to the network for mobile terminals to achieve
real-time glaucoma-assisted detection.

Author Contributions: Conceptualization, Y.C.; data curation, Y.C.; formal analysis, Y.C.; investiga-
tion, Y.C. and Z.L.; methodology, Y.C. and Z.L.; software, Y.C. and Z.L.; validation, Y.C., Y.M. and
Z.L.; visualization, Y.C. and Z.L.; writing—original draft preparation, Y.C. and Z.L.; writing—review
and editing, Y.M. and J.L.; resources, J.L.; supervision, J.L. All authors have read and agreed to the
published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Data Availability Statement: All data are available in the public domain.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Tham, Y.C.; Li, X.; Wong, T.Y.; Quigley, H.A.; Aung, T.; Cheng, C.Y. Global prevalence of glaucoma and projections of glaucoma
burden through 2040: A systematic review and meta-analysis. Ophthalmology 2014, 121, 2081–2090. [CrossRef] [PubMed]
2. Mary, V.S.; Rajsingh, E.B.; Naik, G.R. Retinal fundus image analysis for diagnosis of glaucoma: A comprehensive survey. IEEE
Access 2016, 4, 4327–4354. [CrossRef]
3. Chen, N.; Lv, X. Research on segmentation model of optic disc and optic cup in fundus. BMC Ophthalmol. 2024, 24, 273. [CrossRef]
[PubMed]
4. Zhu, Q.; Chen, X.; Meng, Q.; Song, J.; Luo, G.; Wang, M.; Shi, F.; Chen, Z.; Xiang, D.; Pan, L.; et al. GDCSeg-Net: General optic disc
and cup segmentation network for multi-device fundus images. Biomed. Opt. Express 2021, 12, 6529–6544. [CrossRef] [PubMed]
5. Chen, C.; Wang, G. IOSUDA: An unsupervised domain adaptation with input and output space alignment for joint optic disc
and cup segmentation. Appl. Intell. 2021, 51, 3880–3898. [CrossRef]
6. Fu, H.; Cheng, J.; Xu, Y.; Wong, D.W.K.; Liu, J.; Cao, X. Joint optic disc and cup segmentation based on multi-label deep network
and polar transformation. IEEE Trans. Med. Imaging 2018, 37, 1597–1605. [CrossRef] [PubMed]
7. Al-Bander, B.; Williams, B.M.; Al-Nuaimy, W.; Al-Taee, M.A.; Pratt, H.; Zheng, Y. Dense fully convolutional segmentation of the
optic disc and cup in colour fundus for glaucoma diagnosis. Symmetry 2018, 10, 87. [CrossRef]
8. Qin, P.; Wang, L.; Lv, H. Optic disc and cup segmentation based on deep learning. In Proceedings of the 2019 IEEE 3rd Information
Technology, Networking, Electronic and Automation Control Conference (ITNEC), Chengdu, China, 15–17 March 2019; IEEE:
Piscataway, NJ, USA, 2019; pp. 1835–1840.
9. Zhang, Y.; Cai, X.; Zhang, Y.; Kang, H.; Ji, X.; Yuan, X. TAU: Transferable Attention U-Net for optic disc and cup segmentation.
Knowl.-Based Syst. 2021, 213, 106668. [CrossRef]
10. Mehmood, M.; Naveed, K.; Khan, H.A.; Naqvi, S.S. EDDense-Net: Fully Dense Encoder Decoder Network for Joint Segmentation
of Optic Cup and Disc. arXiv 2023, arXiv:2308.10192.
11. Tadisetty, S.; Chodavarapu, R.; Jin, R.; Clements, R.J.; Yu, M. Identifying the edges of the optic cup and the optic disc in glaucoma
patients by segmentation. Sensors 2023, 23, 4668. [CrossRef] [PubMed]
12. Jiang, L.; Tang, X.; You, S.; Liu, S.; Ji, Y. BEAC-Net: Boundary-Enhanced Adaptive Context Network for Optic Disk and Optic
Cup Segmentation. Appl. Sci. 2023, 13, 10244. [CrossRef]
13. Yu, J.; Chen, N.; Li, J.; Xue, L.; Chen, R.; Yang, C.; Xue, L.; Li, Z.; Wei, L. LC-MANet: Location-constrained joint optic disc and cup
segmentation via multiplex aggregation network. Comput. Electr. Eng. 2024, 118, 109423. [CrossRef]
14. Chen, L.C.; Zhu, Y.; Papandreou, G.; Schroff, F.; Adam, H. Encoder-decoder with atrous separable convolution for semantic image
segmentation. In Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, 8–14 September 2018;
pp. 801–818.
15. He, K.; Zhang, X.; Ren, S.; Sun, J. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 26 June–1 July 2016; pp. 770–778.
16. Howard, A.G.; Zhu, M.; Chen, B.; Kalenichenko, D.; Wang, W.; Weyand, T.; Andreetto, M.; Adam, H. Mobilenets: Efficient
convolutional neural networks for mobile vision applications. arXiv 2017, arXiv:1704.04861.
17. Sandler, M.; Howard, A.; Zhu, M.; Zhmoginov, A.; Chen, L.C. Mobilenetv2: Inverted residuals and linear bottlenecks. In
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–23 June 2018;
pp. 4510–4520.
18. Howard, A.; Sandler, M.; Chu, G.; Chen, L.C.; Chen, B.; Tan, M.; Wang, W.; Zhu, Y.; Pang, R.; Vasudevan, V.; et al. Searching
for mobilenetv3. In Proceedings of the IEEE/CVF International Conference on Computer Vision, Seoul, Republic of Korea, 27
October–2 November 2019; pp. 1314–1324.
Biomimetics 2024, 9, 637 13 of 13

19. Ronneberger, O.; Fischer, P.; Brox, T. U-net: Convolutional networks for biomedical image segmentation. In Proceedings of the
International Conference on Medical Image Computing and Computer-Assisted Intervention, Munich, Germany, 5–9 October
2015; Springer: Berlin/Heidelberg, Germany, 2015; pp. 234–241.
20. Wang, S.; Yu, L.; Yang, X.; Fu, C.W.; Heng, P.A. Patch-based output space adversarial learning for joint optic disc and cup
segmentation. IEEE Trans. Med. Imaging 2019, 38, 2485–2495. [CrossRef] [PubMed]
21. Murugesan, B.; Sarveswaran, K.; Shankaranarayana, S.M.; Ram, K.; Joseph, J.; Sivaprakasam, M. Psi-Net: Shape and boundary
aware joint multi-task deep network for medical image segmentation. In Proceedings of the 2019 41st Annual International
Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Berlin, Germany, 23–27 July 2019; IEEE: Piscataway,
NJ, USA, 2019; pp. 7223–7226.
22. Wang, S.; Yu, L.; Li, K.; Yang, X.; Fu, C.W.; Heng, P.A. Boundary and entropy-driven adversarial learning for fundus image
segmentation. In Proceedings of the Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd
International Conference, Shenzhen, China, 13–17 October 2019, Proceedings, Part I 22; Springer: Berlin/Heidelberg, Germany,
2019; pp. 102–110.
23. Son, J.; Park, S.J.; Jung, K.H. Towards accurate segmentation of retinal vessels and the optic disc in fundoscopic images with
generative adversarial networks. J. Digit. Imaging 2019, 32, 499–512. [CrossRef] [PubMed]
24. Luo, L.; Xue, D.; Pan, F.; Feng, X. Joint optic disc and optic cup segmentation based on boundary prior and adversarial learning.
Int. J. Comput. Assist. Radiol. Surg. 2021, 16, 905–914. [CrossRef] [PubMed]
25. Sevastopolsky, A. Optic disc and cup segmentation methods for glaucoma detection with modification of U-Net convolutional
neural network. Pattern Recognit. Image Anal. 2017, 27, 618–624. [CrossRef]
26. Zilly, J.; Buhmann, J.M.; Mahapatra, D. Glaucoma detection using entropy sampling and ensemble learning for automatic optic
cup and disc segmentation. Comput. Med. Imaging Graph. 2017, 55, 28–41. [CrossRef] [PubMed]
27. Gu, Z.; Cheng, J.; Fu, H.; Zhou, K.; Hao, H.; Zhao, Y.; Zhang, T.; Gao, S.; Liu, J. Ce-net: Context encoder network for 2d medical
image segmentation. IEEE Trans. Med. Imaging 2019, 38, 2281–2292. [CrossRef] [PubMed]
28. Tabassum, M.; Khan, T.M.; Arsalan, M.; Naqvi, S.S.; Ahmed, M.; Madni, H.A.; Mirza, J. CDED-Net: Joint segmentation of optic
disc and optic cup for glaucoma screening. IEEE Access 2020, 8, 102733–102747. [CrossRef]
29. Shankaranarayana, S.M.; Ram, K.; Mitra, K.; Sivaprakasam, M. Fully convolutional networks for monocular retinal depth
estimation and optic disc-cup segmentation. IEEE J. Biomed. Health Inform. 2019, 23, 1417–1426. [CrossRef] [PubMed]
30. Orlando, J.I.; Fu, H.; Breda, J.B.; Keer, K.V.; Bathula, D.R.; Diaz-Pinto, A.; Fang, R.; Heng, P.A.; Kim, J.; Lee, J.H. REFUGE
Challenge: A Unified Framework for Evaluating Automated Methods for Glaucoma Assessment from Fundus Photographs. Med.
Image Anal. 2019, 59, 101570. [CrossRef] [PubMed]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.