Detection of shallow anterior chamber depth from two-dimensional anterior segment photographs using deep learning.

BACKGROUND: The purpose of this study was to implement and evaluate a deep learning (DL) approach for automatically detecting shallow anterior chamber depth (ACD) from two-dimensional (2D) overview anterior segment photographs. METHODS: We trained a DL model using a dataset of anterior segment photographs collected from Shanghai Aier Eye Hospital from June 2018 to December 2019. A Pentacam HR system was used to capture a 2D overview eye image and measure the ACD. Shallow ACD was defined as ACD less than 2.4 mm. The DL model was evaluated by a five-fold cross-validation test in a hold-out testing dataset. We also evaluated the DL model by testing it against two glaucoma specialists. The performance of the DL model was calculated by metrics, including accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). RESULTS: A total of 3753 photographs (1720 shallow AC and 2033 deep AC images) were assigned to the training dataset, and 1302 photographs (509 shallow AC and 793 deep AC images) were held out for two internal testing datasets. In detecting shallow ACD in the internal hold-out testing dataset, the DL model achieved an AUC of 0.86 (95% CI, 0.83-0.90) with 80% sensitivity and 79% specificity. In the same testing dataset, the DL model also achieved better performance than the two glaucoma specialists (accuracy of 80% vs. accuracy of 74 and 69%). CONCLUSIONS: We proposed a high-performing DL model to automatically detect shallow ACD from overview anterior segment photographs. Our DL model has potential applications in detecting and monitoring shallow ACD in the real world. TRIAL REGISTRATION: http://clinicaltrials.gov , NCT04340635 , retrospectively registered on 29 March 2020.

Ocular Anatomical and Functional Characteristics in Anisometropic Chinese Children.

SIGNIFICANCE: This research found that anterior and posterior biometrics differ in many aspects between fellow eyes of anisometropic children. This might shed light on the mechanisms underlying the onset and progression of anisometropia and myopia. PURPOSE: This study aimed to investigate the ocular biometric parameters, peripheral refraction, and accommodative lag of fellow eyes in anisometropic children. METHODS: Anisometropic children were recruited. Axial length (AL), vitreous chamber depth (VCD), central corneal thickness, anterior chamber depth (ACD), lens thickness (LT), simulated K readings, central and peripheral refractive errors, and accommodative lag were measured in both eyes. The subfoveal choroidal thickness, average choroidal thickness, and choroid vessel density of the 6 × 6-mm macular area were measured by optical coherence tomography. RESULTS: Thirty-two children aged 11.1 ± 1.7 years were enrolled. The average degree of anisometropia was 2.49 ± 0.88 D. The AL, VCD, ACD, and simulated K reading values were significantly larger in the more myopic eyes, whereas the LT value was significantly smaller. Subfoveal choroidal thickness (P = .001) and average choroidal thickness (P = .02) were smaller in the more myopic eyes than in the contralateral eyes, whereas choroid vessel density (P = .03) was larger. The amount of anisometropia had a significant positive correlation with the difference in AL (r = 0.869, P < .001), VCD (r = 0.853, P < .001), and ACD (r = 0.591, P < .001) and a negative correlation with the difference in LT (r = -0.457, P = .009). CONCLUSIONS: Ocular biometrics differ in many aspects between the fellow eyes of anisometropic Chinese children, and the difference is correlated with the degree of anisometropia.

Universal digital filtering for denoising volumetric retinal OCT and OCT angiography in 3D shearlet domain.

Retinal optical coherence tomography (OCT) and OCT angiography (OCTA) suffer from the degeneration of image quality due to speckle noise and bulk-motion noise, respectively. Because the cross-sectional retina has distinct features in OCT and OCTA B-scans, existing digital filters that can denoise OCT efficiently are unable to handle the bulk-motion noise in OCTA. In this Letter, we propose a universal digital filtering approach that is capable of minimizing both types of noise. Considering that the retinal capillaries in OCTA are hard to differentiate in B-scans while having distinct curvilinear structures in 3D volumes, we decompose the volumetric OCT and OCTA data with 3D shearlets, thus efficiently separating the retinal tissue and vessels from the noise in this transform domain. Compared with wavelets and curvelets, the shearlets provide better representation of the layer edges in OCT and the vasculature in OCTA. Qualitative and quantitative results show the proposed method outperforms the state-of-the-art OCT and OCTA denoising methods. Also, the superiority of 3D denoising is demonstrated by comparing the 3D shearlet filtering with its 2D counterpart.

Degeneracy-analogous femtosecond dual-wavelength optical parametric oscillator at non-degenerate wavelengths.

Synchronously pumped optical parametric oscillator (OPO) at degeneracy is ideal for generating ultrafast laser pulses. Normally, however, group velocity mismatch (GVM) is ubiquitous among the interacting pulses at widely separated wavelengths. A versatile quasi-phase-matching (QPM) technique is proposed for temporal synchronizing of the signal and idler pulses relied on a less common Type-II QPM (oe-o interaction). The proposed group-velocity regulation technology is advantageous to constructing a degeneracy-analogous femtosecond OPO for dual-wavelength operation. Qualitative prediction for the proposed design is conducted based on a commercial femtosecond pump source at 1064 nm while the signal/idler wavelengths are 3.2 µm and 1.59 µm respectively. Compared with the conventional Type-0 QPM based counterpart (ee-e interaction), the uncompensated temporal distortion caused by temporal walk-off is strongly suppressed while the idler spectrum gets significantly broader. The versatility of the proposed scheme is also clearly demonstrated by its fairly stable performance within a broad tuning range of 2.9-3.5 µm and 1.68-1.53 µm. The demonstrated configuration might be promising for synchronously obtaining dual-wavelength ultrafast pulses with higher spectral and temporal qualities.

Versatile backconversion-inhibited broadband optical parametric amplification based on an idler-separated QPM configuration.

Conversion efficiency and phase-matching (PM) bandwidth are both critical issues for broadband parametric processes. In some sense, they determine the highest peak power achieved via the optical parametric amplification. In this Letter, a versatile idler-separated quasi-phase matching scheme capable of both backconversion circumvention and ultra-broadband PM is presented. Full-dimensional spatial-temporal simulations for the typical optical parametric chirped pulse amplification processes at 800 nm and 3.4 µm were presented in detail. By virtue of the broad PM bandwidth on account of the non-collinear PM configuration, the backconversion circumvention on account of the idler-separated design, and the walk-off self-compensation on account of the symmetrical tilting grating patterns, significantly improved gain bandwidth, extremely high conversion efficiency, and a well-preserved beam profile are simultaneously achieved. Compared with the collinear configuration, the peak power can be potentially enhanced by 5-10 times under the same operation circumstances.

Spectrum regulation for mid-infrared ultrafast pulses via a time-synchronization aperiodically poled LiNbO3.

Restricted to temporal separation during the coupled-waves interaction, aperiodically quasi-phase-matching (QPM) nonlinear crystals are primarily implemented for prechirped pulses, showing limited applications in ultrafast temporal scale. Under the proposed time-synchronization framework, pump and signal waves travel with identical group-velocity, which permits sustaining energy transfer in long aperiodically poled LiNbO3 crystals (APPLN) even with ultrafast pulse duration. With the help of this structure, adiabatic frequency conversion shows extra advantages compared with the common cases, which enables lower stretching ratio and smoother gain spectrum. Focusing on the typical mid-infrared wavelength of ~3 µm, we numerically study the potential performance of APPLN with chirp-free ultrabroad interacting waves. In contrast to the spectral shift and conversion efficiency degradation presented by its traditional Type-0 QPM counterpart, the proposed design demonstrated impressive ability to obtain arbitrary spectrum via a simple femtosecond OPA/OPO. Peculiarly, the QPM chirp rate sign plays a significant role to the output spectrum, and a positive chirp rate is preferential in delivering a bandwidth-controllable spectrum. The proposed design provides a promising technical route to achieve spectrum manipulation in ultrafast temporal scale.

High power tandem-pumped thulium-doped fiber laser.

We propose a cascaded tandem pumping technique and show its high power and high efficient operation in the 2-µm wavelength region, opening up a new way to scale the output power of the 2-µm fiber laser to new levels (e.g. 10 kW). Using a 1942 nm Tm(3+) fiber laser as the pump source with the co- (counter-) propagating configuration, the 2020 nm Tm(3+) fiber laser generates 34.68 W (35.15W) of output power with 84.4% (86.3%) optical-to-optical efficiency and 91.7% (92.4%) slope efficiency, with respect to launched pump power. It provides the highest slope efficiency reported for 2-µm Tm(3+)-doped fiber lasers, and the highest output power for all-fiber tandem-pumped 2-µm fiber oscillators. This system fulfills the complete structure of the proposed cascaded tandem pumping technique in the 2-µm wavelength region (~1900 nm → ~1940 nm → ~2020 nm). Numerical analysis is also carried out to show the power scaling capability and efficiency of the cascaded tandem pumping technique.

High signal-to-noise ratio, single-frequency 2 µm Brillouin fiber laser.

A high optical signal-to-noise ratio (OSNR) single-frequency 2 µm Brillouin fiber laser (BFL) with watt-level output and high transfer efficiency is demonstrated for the first time to the best of our knowledge. The Brillouin pump is constructed with a two-stage thulium-doped fiber amplifier (TDFA) seeded by a 2 µm laser diode, providing 4.02 W average power with 1 MHz linewidth. Using an optimized length of 14 m for the Brillouin ring cavity, the BFL works stably in single-mode region with 8 kHz linewidth because of the linewidth narrowing effect. The transfer efficiency is 51% with 1.08 W output power and 62 dB OSNR for 3.22 W pump power.

Robust superhydrophobic cotton fabric based on dual-sized silica particles with self-healing nature.

Improving the durability of wear-resistant superhydrophobic surfaces is crucial for their practical use. To tackle this, research is now delving into self-healing superhydrophobic surfaces. In our study, we developed superhydrophobic cotton fabrics by embedding nano-silica particles, micro-silica powder, and polydimethylsiloxane (PDMS) using a dipping method. This innovative design grants the SiO2/PDMS cotton fabric remarkable superhydrophobicity, reflected by a water contact angle of 155°. Moreover, the PDMS was stored in the amorphous areas of cellulose of cotton fabrics, attaching to the fiber surface and playing a role in connecting micro-blocks and nano-particles. This causes a self-diffusion of PDMS molecules in these fabrics, allowing the surface to regain its superhydrophobicity even after abrasion damage. Impressively, this self-healing property can be renewed at least 8 times, showcasing the fabric's resilience. Moreover, these superhydrophobic cotton fabrics exhibit outstanding self-cleaning abilities and repel various substances such as blood, milk, cola, and tea. This resilience, coupled with its simplicity, low cost-effectiveness, and eco-friendliness, makes this coating highly promising for applications across construction, chemical, and medical fields. Our study also delves into understanding the self-healing mechanism of the SiO2/PDMS cotton fabric, offering insights into their long-term performance and potential advancements in this field.

Towards ultra-low-cost smartphone microscopy.

The outbreak of COVID-19 exposed the inadequacy of our technical tools for home health surveillance, and recent studies have shown the potential of smartphones as a universal optical microscopic imaging platform for such applications. However, most of them use laboratory-grade optomechanical components and transmitted illuminations to ensure focus tuning capability and imaging quality, which keeps the cost of the equipment high. Here, we propose an ultra-low-cost solution for smartphone microscopy. To realize focus tunability, we designed a seesaw-like structure capable of converting large displacements on one side into small displacements on the other (reduced to ∼9.1%), which leverages the intrinsic flexibility of 3D printing materials. We achieved a focus-tuning accuracy of ∼5 𝜇m, which is 40 times higher than the machining accuracy of the 3D-printed lens holder itself. For microscopic imaging, we used an off-the-shelf smartphone camera lens as the objective and the built-in flashlight as the illumination. To compensate for the resulting image quality degradation, we developed a learning-based image enhancement method. We used the CycleGAN architecture to establish the mapping from smartphone microscope images to benchtop microscope images without pairing. We verified the imaging performance on different biomedical samples. Except for the smartphone, we kept the full costs of the device under 4 USD. We think these efforts to lower the costs of smartphone microscopes will benefit their applications in various scenarios, such as point-of-care testing, on-site diagnosis, and home health surveillance. RESEARCH HIGHLIGHTS: We propose a solution for ultra-low-cost smartphone microscopy. Utilizing the flexibility of 3D-printed material, we can achieve focusing accuracy of ∼5 𝜇m. Such a low-cost device will benefit point-of-care diagnosis and home health surveillance.

Learning-based distortion correction enables proximal-scanning endoscopic OCT elastography.

Proximal rotary scanning is predominantly used in the clinical practice of endoscopic and intravascular OCT, mainly because of the much lower manufacturing cost of the probe compared to distal scanning. However, proximal scanning causes severe beam stability issues (also known as non-uniform rotational distortion, NURD), which hinders the extension of its applications to functional imaging, such as OCT elastography (OCE). In this work, we demonstrate the abilities of learning-based NURD correction methods to enable the imaging stability required for intensity-based OCE. Compared with the previous learning-based NURD correction methods that use pseudo distortion vectors for model training, we propose a method to extract real distortion vectors from a specific endoscopic OCT system, and validate its superiority in accuracy under both convolutional-neural-network- and transformer-based learning architectures. We further verify its effectiveness in elastography calculations (digital image correlation and optical flow) and the advantages of our method over other NURD correction methods. Using the air pressure of a balloon catheter as a mechanical stimulus, our proximal-scanning endoscopic OCE could effectively differentiate between areas of varying stiffness of atherosclerotic vascular phantoms. Compared with the existing endoscopic OCE methods that measure only in the radial direction, our method could achieve 2D displacement/strain distribution in both radial and circumferential directions.

Cross-attention learning enables real-time nonuniform rotational distortion correction in OCT.

Nonuniform rotational distortion (NURD) correction is vital for endoscopic optical coherence tomography (OCT) imaging and its functional extensions, such as angiography and elastography. Current NURD correction methods require time-consuming feature tracking/registration or cross-correlation calculations and thus sacrifice temporal resolution. Here we propose a cross-attention learning method for the NURD correction in OCT. Our method is inspired by the recent success of the self-attention mechanism in natural language processing and computer vision. By leveraging its ability to model long-range dependencies, we can directly obtain the spatial correlation between OCT A-lines at any distance, thus accelerating the NURD correction. We develop an end-to-end stacked cross-attention network and design three types of optimization constraints. We compare our method with two traditional feature-based methods and a CNN-based method on two publicly-available endoscopic OCT datasets. We further verify the NURD correction performance of our method on 3D stent reconstruction using a home-built endoscopic OCT system. Our method achieves a ∼3 × speedup to real time (26 ± 3 fps), and superior correction performance.

Generative adversarial networks synthetic optical coherence tomography images as an education tool for image diagnosis of macular diseases: a randomized trial.

Purpose: This study aimed to evaluate the effectiveness of generative adversarial networks (GANs) in creating synthetic OCT images as an educational tool for teaching image diagnosis of macular diseases to medical students and ophthalmic residents. Methods: In this randomized trial, 20 fifth-year medical students and 20 ophthalmic residents were enrolled and randomly assigned (1:1 allocation) into Group real OCT and Group GANs OCT. All participants had a pretest to assess their educational background, followed by a 30-min smartphone-based education program using GANs or real OCT images for macular disease recognition training. Two additional tests were scheduled: one 5 min after the training to assess short-term performance, and another 1 week later to assess long-term performance. Scores and time consumption were recorded and compared. After all the tests, participants completed an anonymous subjective questionnaire. Results: Group GANs OCT scores increased from 80.0 (46.0 to 85.5) to 92.0 (81.0 to 95.5) 5 min after training (p < 0.001) and 92.30 ± 5.36 1 week after training (p < 0.001). Similarly, Group real OCT scores increased from 66.00 ± 19.52 to 92.90 ± 5.71 (p < 0.001), respectively. When compared between two groups, no statistically significant difference was found in test scores, score improvements, or time consumption. After training, medical students had a significantly higher score improvement than residents (p < 0.001). Conclusion: The education tool using synthetic OCT images had a similar educational ability compared to that using real OCT images, which improved the interpretation ability of ophthalmic residents and medical students in both short-term and long-term performances. The smartphone-based educational tool could be widely promoted for educational applications.Clinical trial registration: https://www.chictr.org.cn, Chinese Clinical Trial Registry [No. ChiCTR 2100053195].

Annotation-efficient learning for OCT segmentation.

Deep learning has been successfully applied to OCT segmentation. However, for data from different manufacturers and imaging protocols, and for different regions of interest (ROIs), it requires laborious and time-consuming data annotation and training, which is undesirable in many scenarios, such as surgical navigation and multi-center clinical trials. Here we propose an annotation-efficient learning method for OCT segmentation that could significantly reduce annotation costs. Leveraging self-supervised generative learning, we train a Transformer-based model to learn the OCT imagery. Then we connect the trained Transformer-based encoder to a CNN-based decoder, to learn the dense pixel-wise prediction in OCT segmentation. These training phases use open-access data and thus incur no annotation costs, and the pre-trained model can be adapted to different data and ROIs without re-training. Based on the greedy approximation for the k-center problem, we also introduce an algorithm for the selective annotation of the target data. We verified our method on publicly-available and private OCT datasets. Compared to the widely-used U-Net model with 100% training data, our method only requires ∼10% of the data for achieving the same segmentation accuracy, and it speeds the training up to ∼3.5 times. Furthermore, our proposed method outperforms other potential strategies that could improve annotation efficiency. We think this emphasis on learning efficiency may help improve the intelligence and application penetration of OCT-based technologies.

PdCu nanoalloy decorated photocatalysts for efficient and selective oxidative coupling of methane in flow reactors.

Methane activation by photocatalysis is one of the promising sustainable technologies for chemical synthesis. However, the current efficiency and stability of the process are moderate. Herein, a PdCu nanoalloy (~2.3 nm) was decorated on TiO2, which works for the efficient, stable, and selective photocatalytic oxidative coupling of methane at room temperature. A high methane conversion rate of 2480 µmol g-1 h-1 to C2 with an apparent quantum efficiency of ~8.4% has been achieved. More importantly, the photocatalyst exhibits the turnover frequency and turnover number of 116 h-1 and 12,642 with respect to PdCu, representing a record among all the photocatalytic processes (λ > 300 nm) operated at room temperature, together with a long stability of over 112 hours. The nanoalloy works as a hole acceptor, in which Pd softens and weakens C-H bond in methane and Cu decreases the adsorption energy of C2 products, leading to the high efficiency and long-time stability.

A Novel Tumor Progression Prediction Method for Multimode Ablation Treatment.

OBJECTIVE: The multimode ablation of liver cancer, which uses radio-frequency heating after a pre-freezing process to treat the tumor, has shown significantly improved therapeutic effects and enhanced anti-tumor immune response. Unlike open surgery, the ablated lesions remain in the body after treatment, so it is critical to assess the immediate outcome and to monitor disease status over time. Here we propose a novel tumor progression prediction method for simultaneous postoperative evaluation and prognosis analysis. METHODS: We propose to leverage the intraoperative therapeutic information extracted from thermal dose distribution. For tumors with specific sensitivity reflected in medical images, different thermal doses implicitly indicate the degree of instant damage and long-term inhibition excited under specific ablation energy. We further propose a survival analysis framework for the multimode ablation treatment. It extracts carefully designed features from clinical, preoperative, intraoperative, and postoperative data, then uses random survival forest for feature selection and deep neural networks for survival prediction. RESULTS: We evaluated the proposed methods using clinical data. The results show that our method outperforms the state-of-the-art survival analysis methods with a C-index of 0.855±0.090. The thermal dose information contributes significantly to the prediction accuracy by taking up 21.7% of the overall feature importance. CONCLUSION: The proposed methods have been demonstrated to be a powerful tool in tumor progression prediction of multimode ablation therapy. SIGNIFICANCE: This kind of data-driven prognosis analysis may benefit personalized medicine and simplify the follow-up process.

Memorizing Structure-Texture Correspondence for Image Anomaly Detection.

This work focuses on image anomaly detection by leveraging only normal images in the training phase. Most previous methods tackle anomaly detection by reconstructing the input images with an autoencoder (AE)-based model, and an underlying assumption is that the reconstruction errors for the normal images are small, and those for the abnormal images are large. However, these AE-based methods, sometimes, even reconstruct the anomalies well; consequently, they are less sensitive to anomalies. To conquer this issue, we propose to reconstruct the image by leveraging the structure-texture correspondence. Specifically, we observe that, usually, for normal images, the texture can be inferred from its corresponding structure (e.g., the blood vessels in the fundus image and the structured anatomy in optical coherence tomography image), while it is hard to infer the texture from a destroyed structure for the abnormal images. Therefore, a structure-texture correspondence memory (STCM) module is proposed to reconstruct image texture from its structure, where a memory mechanism is used to characterize the mapping from the normal structure to its corresponding normal texture. As the correspondence between destroyed structure and texture cannot be characterized by the memory, the abnormal images would have a larger reconstruction error, facilitating anomaly detection. In this work, we utilize two kinds of complementary structures (i.e., the semantic structure with human-labeled category information and the low-level structure with abundant details), which are extracted by two structure extractors. The reconstructions from the two kinds of structures are fused together by a learned attention weight to get the final reconstructed image. We further feed the reconstructed image into the two aforementioned structure extractors to extract structures. On the one hand, constraining the consistency between the structures extracted from the original input and that from the reconstructed image would regularize the network training; on the other hand, the error between the structures extracted from the original input and that from the reconstructed image can also be used as a supplement measurement to identify the anomaly. Extensive experiments validate the effectiveness of our method for image anomaly detection on both industrial inspection images and medical images.

Proxy-Bridged Image Reconstruction Network for Anomaly Detection in Medical Images.

Anomaly detection in medical images refers to the identification of abnormal images with only normal images in the training set. Most existing methods solve this problem with a self-reconstruction framework, which tends to learn an identity mapping and reduces the sensitivity to anomalies. To mitigate this problem, in this paper, we propose a novel Proxy-bridged Image Reconstruction Network (ProxyAno) for anomaly detection in medical images. Specifically, we use an intermediate proxy to bridge the input image and the reconstructed image. We study different proxy types, and we find that the superpixel-image (SI) is the best one. We set all pixels' intensities within each superpixel as their average intensity, and denote this image as SI. The proposed ProxyAno consists of two modules, a Proxy Extraction Module and an Image Reconstruction Module. In the Proxy Extraction Module, a memory is introduced to memorize the feature correspondence for normal image to its corresponding SI, while the memorized correspondence does not apply to the abnormal images, which leads to the information loss for abnormal image and facilitates the anomaly detection. In the Image Reconstruction Module, we map an SI to its reconstructed image. Further, we crop a patch from the image and paste it on the normal SI to mimic the anomalies, and enforce the network to reconstruct the normal image even with the pseudo abnormal SI. In this way, our network enlarges the reconstruction error for anomalies. Extensive experiments on brain MR images, retinal OCT images and retinal fundus images verify the effectiveness of our method for both image-level and pixel-level anomaly detection.

Development and Clinical Validation of Semi-Supervised Generative Adversarial Networks for Detection of Retinal Disorders in Optical Coherence Tomography Images Using Small Dataset.

PURPOSE: To develop and test semi-supervised generative adversarial networks (GANs) that detect retinal disorders on optical coherence tomography (OCT) images using a small-labeled dataset. METHODS: From a public database, we randomly chose a small supervised dataset with 400 OCT images (100 choroidal neovascularization, 100 diabetic macular edema, 100 drusen, and 100 normal) and assigned all other OCT images to unsupervised dataset (107,912 images without labeling). We adopted a semi-supervised GAN and a supervised deep learning (DL) model for automatically detecting retinal disorders from OCT images. The performance of the 2 models was compared in 3 testing datasets with different OCT devices. The evaluation metrics included accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curves. RESULTS: The local validation dataset included 1000 images with 250 from each category. The independent clinical dataset included 366 OCT images using Cirrus OCT Shanghai Shibei Hospital and 511 OCT images using RTVue OCT from Xinhua Hospital respectively. The semi-supervised GANs classifier achieved better accuracy than supervised DL model (0.91 vs 0.86 for local cell validation dataset, 0.91 vs 0.86 in the Shanghai Shibei Hospital testing dataset, and 0.93 vs 0.92 in Xinhua Hospital testing dataset). For detecting urgent referrals (choroidal neo-vascularization and diabetic macular edema) from nonurgent referrals (drusen and normal) on OCT images, the semi-supervised GANs classifier also achieved better area under the receiver operating characteristic curves than supervised DL model (0.99 vs 0.97, 0.97 vs 0.96, and 0.99 vs 0.99, respectively). CONCLUSIONS: A semi-supervised GAN can achieve better performance than that of a supervised DL model when the labeled dataset is limited. The current study offers utility to various research and clinical studies using DL with relatively small datasets. Semi-supervised GANs can detect retinal disorders from OCT images using relatively small dataset.

Noise reduction by adaptive-SIN filtering for retinal OCT images.

Optical coherence tomography (OCT) images is widely used in ophthalmic examination, but their qualities are often affected by noises. Shearlet transform has shown its effectiveness in removing image noises because of its edge-preserving property and directional sensitivity. In the paper, we propose an adaptive denoising algorithm for OCT images. The OCT noise is closer to the Poisson distribution than the Gaussian distribution, and shearlet transform assumes additive white Gaussian noise. We hence propose a square-root transform to redistribute the OCT noise. Different manufacturers and differences between imaging objects may influence the observed noise characteristics, which make predefined thresholding scheme ineffective. We propose an adaptive 3D shearlet image filter with noise-redistribution (adaptive-SIN) scheme for OCT images. The proposed adaptive-SIN is evaluated on three benchmark datasets using quantitative evaluation metrics and subjective visual inspection. Compared with other algorithms, the proposed algorithm better removes noise in OCT images and better preserves image details, significantly outperforming in terms of both quantitative evaluation and visual inspection. The proposed algorithm effectively transforms the Poisson noise to Gaussian noise so that the subsequent shearlet transform could optimally remove the noise. The proposed adaptive thresholding scheme optimally adapts to various noise conditions and hence better remove the noise. The comparison experimental results on three benchmark datasets against 8 compared algorithms demonstrate the effectiveness of the proposed approach in removing OCT noise.