A new detection model of microaneurysms based on improved FC-DenseNet

Diabetic retinopathy (DR) is a frequent vascular complication of diabetes mellitus and remains a leading cause of vision loss worldwide. Microaneurysm (MA) is usually the first symptom of DR that leads to blood leakage in the retina. Periodic detection of MAs will facilitate early detection of DR and reduction of vision injury. In this study, we proposed a novel model for the detection of MAs in fluorescein fundus angiography (FFA) images based on the improved FC-DenseNet, MAs-FC-DenseNet. FFA images were pre-processed by the Histogram Stretching and Gaussian Filtering algorithm to improve the quality of FFA images. Then, MA regions were detected by the improved FC-DenseNet. MAs-FC-DenseNet was compared against other FC-DenseNet models (FC-DenseNet56 and FC-DenseNet67) or the end-to-end models (DeeplabV3+ and PSPNet) to evaluate the detection performance of MAs. The result suggested that MAs-FC-DenseNet had higher values of evaluation metrics than other models, including pixel accuracy (PA), mean pixel accuracy (MPA), precision (Pre), recall (Re), F1-score (F1), and mean intersection over union (MIoU). Moreover, MA detection performance for MAs-FC-DenseNet was very close to the ground truth. Taken together, MAs-FC-DenseNet is a reliable model for rapid and accurate detection of MAs, which would be used for mass screening of DR patients.

Retinal microaneurysms (MAs) are defined as the small swelling of tiny blood vessels, which mainly locate in the inner nuclear layer and deep capillary layer 1 . MAs often occur as the early clinical signs of retinal diseases, such as diabetic retinopathy (DR) and retinal vein occlusions. The number and turnover of retinal MAs are considered as the indicators to assess the presence, severity, and progression risk of retinopathy. Thus, periodic detection of MAs is required for the early diagnosis of retinopathy. MAs can be identified by several imaging technologies, including color fundus photography, fundus fluorescein angiography (FFA), and optical coherence tomography angiography (OCTA). Clinically, FFA is well-recognized as an important standard to visualize retinal vasculature and is routinely used to describe the subtle vascular alterations 2 .
FFA is highly sensitive and demonstrates MAs as the hyperfluorescent dots in the early phase. It is an important imaging modality, which can capture images after the intravenous injection of fluorescein dye 3 . With the increasing amount of FFA images that require for analysis, manual detection and quantification of MAs have become the labor-intensive and time-consuming jobs 4 . In addition, manual detection of MAs is subjective and error-prone, which may cause poor reproducibility 5 . Thus, an automated detection method is urgently required for the accurate detection of MAs in FFA images.
Recently, the development of MA detection methods have become a hot topic in the field of ophthalmic study. Some models based on the neural network have been used for MA detection. CNN and ResNet could obtain higher-level features from the upper layer and give up the features of the lower layer, but these model may loss parts of small-size targets 6,7 . DeeplabV3+and PSPNet use spatial pyramid pooling module to further extract contextual information and improve the detection accuracy of small-size targets, but they have misdetection and missed detection problems for MA detection 8,9 . Some improvements on neural network have been used for MA detection. Mazlan    Pre-processing of FFA images by the histogram stretching and Gaussian filtering. FFA images were pre-processed to improve their contrast and reduce image noises. The contrast between MAs and the background was enhanced by the Histogram stretching. Histogram stretching could stretch the values of pixel from 0 to 255 [15][16][17] . The noises of FFA images were then reduced by the Gaussian Filtering. Gaussian Filtering could remove the surrounding noises in the non-uniform retinal images [18][19][20] .

Materials and methods
Histogram Stretching is expressed as where I new is the new transformed image; I max and I min are the largest and smallest possible grey level value in the original image, respectively. G max and G min are the largest and smallest possible grey level value in the transformed image, respectively. Gaussian Filtering is expressed as where σ 2 is the variance of Gaussian Filtering; l is the size of the filter kernel.

MA detection by the improved FC-DenseNet. At this step, MAs were detected by the improved Fully
Convolutional DenseNet (FC-DenseNet) 21 . FC-DenseNet contains the down sampling path for extracting sparse semantic features and the up sampling path for restoring original resolution. The down sampling path consists of dense block (DB) layer and transition down (TD) layer 22 . The up sampling path consists of DB layer and transition up (TU) layer. DB layer is composed of batch normalization (BN) 23 , ReLU 24 , 3 × 3 convolution, and dropout with probability p = 0.2. TD layer is composed of BN, ReLU, 1 × 1 convolution, dropout with probability p = 0.2 and 2 × 2 maximum pooling. TU layer includes 3 × 3 transposed convolution with stride 2. As shown in Fig. 2, the feature maps from the down sampling path were concatenated with the corresponding feature maps in the up sampling path. The connectivity pattern in the up sampling and the down sampling paths were different. In the down sampling path, the input to a dense block was concatenated with its output, leading to a linear growth of the number of feature maps, whereas in the up sampling path, it was not.
(1) Here, the focal loss function is expressed as where p t is probability of correct prediction for different categories. α t and γ ≥ 0 are adjustable hyperparameters, which can be used to control the sharing weight of different samples to the total loss.

Results
The original FFA images, the detection results by MAs-FC-DenseNet, and the ground truth by the clinicians were shown in Fig. 3. To evaluate the detection performance for MAs, three comparison experiments were conducted. Experiment 1 was an ablation experiment. In experiment 2, MAs-FC-DenseNet was compared against other FC-DenseNet models including FC-DenseNet56 and FC-DenseNet67 to compare MA detection performance. In experiment 3, MAs-FC-DenseNet was compared against other end-to-end models, including DeeplabV3+ and PSPNet, to compare MA detection performance. Ablation experiment. Figure 4 and Table 1 showed the comparison result of MA detection performance in the ablation studies, including FC-DenseNet103 with pre-processing (pre-processing + FC-DenseNet103), the improved FC-DenseNet103 without pre-processing (FC-DenseNet103 + Focal loss), and the proposed MAs-FC-DenseNet.
As shown in Fig. 4 and Table 1, the models of FC-DenseNet103 with pre-processing or improved FC-DenseNet103 without pre-processing led to some missed and false detection of MAs. By contrast, MAs-FC-DenseNet reduced the missed and false detection regions and significantly enhanced the values of PA, MPA, Pre, Re, F1, and MIoU.

Evaluation of MA detection performance of MAs-FC-DenseNet by comparing against other FC-DenseNet models. MAs-FC-DenseNet was compared against other FC-DenseNet models, including
FC-DenseNet56 and FC-DenseNet67, to evaluate the detection performance of MAs.
As shown in Fig. 5 and Table 2 Evaluation of MA detection performance of MAs-FC-DenseNet by comparing against other end-to-end models. MAs-FC-DenseNet was compared against other end-to-end models, including Deep-labV3+ and PSPNet, to evaluate the detection performance of MAs.
As shown in Fig. 6 and Table 3, the DeeplabV3+ model could not distinguish the boundaries of MAs and normal regions, which led to some false detection of MAs. As for PSPNet model, some MA regions were missed.

Discussion
Microaneurysm (MA) is recognized as the first symptom of DR that leads to retinal blood injury. Detection of MAs within FFA images facilitates early DR detection and prevents vision loss 30 . However, MA is extremely small and its contrast to the surrounding background is very subtle, which make MA detection challenging. In this study, we proposed a novel model, MAs-FC-DenseNet, for the detection of MAs in FFA images. FFA images were pre-processed to enhance image quality by the Histogram Stretching and Gaussian Filtering. Improved FC-DenseNet model was then used to detect the deep features of MAs and enhance the detection accuracy of MAs. Six metrics were used to evaluate the accuracy for MA detection, including pixel accuracy (PA), mean pixel accuracy (MPA), Precision (Pre), Recall (Re), F1-score (F1), and mean intersection over union (MIoU). The values of these metrics of MAs-FC-DenseNet were significantly greater than that of other deep learning network models, including DeeplabV3+ and PSPNet, which could reach to 99.97%, 94.19%, 88.40%, 89.70%, 88.98%, and 90.14%, respectively. Moreover, the detection results of MAs-FC-DenseNet were very close to the ground truth. Thus, MAs-FC-DenseNet is a suitable model for diabetic retinopathy screening based on MA detection result.
DR is a complex, progressive, and heterogenous ocular disease associated with diabetes duration. It is generally recognized as a vascular disease like other diabetes-related diseases. Signs of DR contain the lesions such as MAs, hemorrhages, and yellowish or bright spots such as hard and soft exudates 31 . This model was designed based on the feature of MAs, but the impacts of other lesions on MA detection were not considered. In addition, DR can be classified as mild non-proliferative DR (NPDR), moderate NPDR, severe NPDR, and proliferative DR (PDR) according to disease severity 32 . Our proposed model did not consider the severity DR. The degree of DR severity should be considered as a contributing factor for MA detection in the future study.
Taken together, this study proposed a two-step model, MAs-FC-DenseNet, for the detection of MAs in FFA images, including pre-processing of FFA images by the Histogram Stretching and Gaussian Filtering and detection of MAs by the improved FC-DenseNet. This model will become a promising method for early diagnosis of diabetic retinopathy with a competitive accuracy. In the future, this model should be improved to embrace more feature-learning capacities, as well as some knowledge regarding retinal geometry and other characters.   www.nature.com/scientificreports/