Binarization Method Research Articles

Application of Deep Learning Methods for Binarization of the Choroid in Optical Coherence Tomography Images.

PurposeThe purpose of this study was to develop a deep learning model for automatic binarization of the choroidal tissue, separating choroidal blood vessels from nonvascular stromal tissue, in optical coherence tomography (OCT) images from healthy young subjects.MethodsOCT images from an observational longitudinal study of 100 children were used for training, validation, and testing of 5 fully semantic networks, which provided a binarized output of the choroid. These outputs were compared with ground truth images, generated from a local binarization technique after manually optimizing the analysis window size for each individual image. The performance was evaluated using accuracy and repeatability metrics. The methods were also compared with a fixed window size local binarization technique, which has been commonly used previously.ResultsThe tested deep learning methods provided a good performance in terms of accuracy and repeatability. With the U-Net and SegNet networks showing >96% accuracy. All methods displayed a high level of repeatability relative to the ground truth. For analysis of the choroidal vascularity index (a commonly used metric derived from the binarized image), SegNet showed the closest agreement with the ground truth and high repeatability. The fixed window size showed a reduced accuracy compared to other methods.ConclusionsFully semantic networks such as U-Net and SegNet displayed excellent performance for the binarization task. These methods provide a useful approach for clinical and research applications of deep learning tools for the binarization of the choroid in OCT images.Translational RelevanceDeep learning models provide a novel, robust solution to automatically binarize the choroidal tissue in OCT images.

BiRe-ID: Binary Neural Network for Efficient Person Re-ID

Person re-identification (Re-ID) has been promoted by the significant success of convolutional neural networks (CNNs). However, the application of such CNN-based Re-ID methods depends on the tremendous consumption of computation and memory resources, which affects its development on resource-limited devices such as next generation AI chips. As a result, CNN binarization has attracted increasing attention, which leads to binary neural networks (BNNs). In this article, we propose a new BNN-based framework for efficient person Re-ID (BiRe-ID). In this work, we discover that the significant performance drop of binarized models for Re-ID task is caused by the degraded representation capacity of kernels and features. To address the issues, we propose the kernel and feature refinement based on generative adversarial learning (KR-GAL and FR-GAL) to enhance the representation capacity of BNNs. We first introduce an adversarial attention mechanism to refine the binarized kernels based on their real-valued counterparts. Specifically, we introduce a scale factor to restore the scale of 1-bit convolution. And we employ an effective generative adversarial learning method to train the attention-aware scale factor. Furthermore, we introduce a self-supervised generative adversarial network to refine the low-level features using the corresponding high-level semantic information. Extensive experiments demonstrate that our BiRe-ID can be effectively implemented on various mainstream backbones for the Re-ID task. In terms of the performance, our BiRe-ID surpasses existing binarization methods by significant margins, at the level even comparable with the real-valued counterparts. For example, on Market-1501, BiRe-ID achieves 64.0% mAP on ResNet-18 backbone, with an impressive 12.51× speedup in theory and 11.75× storage saving. In particular, the KR-GAL and FR-GAL methods show strong generalization on multiple tasks such as Re-ID, image classification, object detection, and 3D point cloud processing.

Evaluation of retinal and choroidal microcirculation in Behçet's disease.

To investigate retinal perfusion by optical coherence tomography (OCT)-angiography and choroidal vascular features using an OCT-based image binarization method in patients with Behçet's disease (BD) without clinical ocular involvement. This study included 38 patients with non-ocular BD and 35 healthy participants. Macular region was evaluated with OCT-angiography (Optovue, Inc., Fremont, CA). A 6.0 × 6.0 mm rectanglescan centred on the fovea was used to record the scans. The enhanced depth imaging OCT scans (Heidelberg Eye Explorer version 1.8.6.0; Heidelberg Engineering) of the macula and peripapillary scans of the optic nerve head were binarized using ImageJ software (National Institutes of Health, Bethesda, MD, USA). The luminal area (LA), stromal area (SA), total choroidal area (TCA), and choroidal thickness (CT) were analysed. The choroidal vascularity index (CVI) was calculated as the ratio of the LA to the TCA. Vessel density values of the superficial and deep macular capillary plexus were similar between the two groups (all p > 0.05). There was no significant difference between the groups in terms of the CT and TCA values in the macula and in all peripapillary sectors (all p > 0.05). However, the macula and temporal, nasal, and inferior sectors of the peripapillary area had significantly lower CVI values in the BD group compared with controls (p = 0.009, p = 0.002, p = 0.010, and p = 0.008, respectively). Retinal microperfusion deficit was not observed in non-ocular Behçet patients. CVI may be a more robust marker than CT to indicate choroidal perfusion. A reduced CVI may suggest subclinical ocular involvement and choroidal ischemia in these patients.

Comparison of choroidal hyperreflective spots on optical coherence tomography images between both eyes of normal subjects.

Advancement of optical coherence tomography (OCT) technology allows for better in vivo visualization of the choroidal architecture, which comprises vessels and stroma. However, most OCT studies using image binarization methods have focused only on choroidal vessels represented by dark pixels. This study aimed to compare the distribution of choroidal hyperreflective spots on swept-source OCT (SS-OCT) images between both eyes of normal subjects. In this observational comparative study, we included SS-OCT images of healthy subjects, which were prospectively obtained to compare images among the devices. SS-OCT images acquired using PLEX Elite 9000 and DRI-OCT Triton were analyzed. En-face OCT images were obtained at five different depth positions of the inner choroid at the macula. The mean reflectivity of the choroidal slabs, the number, total area, and circularity of hyperreflective spots were quantitatively compared between the devices and between both eyes of the same subjects. In 30 eyes of 15 healthy subjects, the mean reflectivity of the choroidal slabs varied with the scan depth on both devices (P<0.001 and P<0.001). Hyperreflective spots were similarly distributed in the images from both devices, but at different depths. The number and area of hyperreflective spots in the second and third layers of the DRI-OCT Triton were positively correlated with those in the fourth and fifth layers of the PLEX Elite 9000, respectively (all P<0.05). The intraclass correlation coefficients (ICC) for the area of hyperreflective spots were excellent for the third slab of the DRI-OCT and the fifth slab of the PLEX Elite (ICC =0.798; 95% CI, -0.576-0.904). The number and area of hyperreflective spots were correlated between both eyes in the third, fourth and fifth layers on DRI-OCT Triton (all P<0.05) and in the first, second, and fifth layers on PLEX Elite 9000 (all P<0.05). Hyperreflective spots in en-face images from two different SS-OCT devices were similarly observed between both eyes of the same person. The distributions of spots between the two eyes of the same person were correlated. These findings suggest that the distribution of hyperreflective spots on the choroid reflects the choroidal characteristics of the subject.

Adaptive Digital Hologram Binarization Method Based on Local Thresholding, Block Division and Error Diffusion.

High-speed optical reconstruction of 3D-scenes can be achieved using digital holography with binary digital micromirror devices (DMD) or a ferroelectric spatial light modulator (fSLM). There are many algorithms for binarizing digital holograms. The most common are methods based on global and local thresholding and error diffusion techniques. In addition, hologram binarization is used in optical encryption, data compression, beam shaping, 3D-displays, nanofabrication, materials characterization, etc. This paper proposes an adaptive binarization method based on a combination of local threshold processing, hologram division into blocks, and error diffusion procedure (the LDE method). The method is applied for binarization of optically recorded and computer-generated digital holograms of flat objects and three-dimensional scenes. The quality of reconstructed images was compared with different methods of error diffusion and thresholding. Image reconstruction quality was up to 22% higher by various metrics than that one for standard binarization methods. The optical hologram reconstruction using DMD confirms the results of the numerical simulations.

Surface topography measurement of microstructures near the lateral resolution limit via coherence scanning interferometry

As a non-destructive method, coherence scanning interferometry is widely utilized for surface topography measurement of step microstructures in optical and integrated circuit fields. The measurements are always implemented near the lateral resolution limit of the system because the critical dimension of these microstructures is at the micron or sub-micron scale. However, in this case, not only is the image of microstructures blurred, but also the diffraction effects distort the measured topography near the step edges. The magnitude of distortion at a certain position appears to be related to its distance from the edges and the duty cycle of the microstructures. Therefore, it is necessary, although difficult, to improve the accuracy of topography measurement for microstructures if their images are blurred owing to the lateral resolution limit. When located far away from the step edges of microstructures, the coherence signals are normal and contain a single envelope with its peak corresponding to the surface height. However, the anomalous coherence signals containing two envelopes appear near the step edges. The peaks of these envelopes correspond to the top and bottom surfaces of step structures, thus distorting the coherence peak detection for topography measurement. Therefore, the primary issue is the determination of the correct envelope in the anomalous coherence signals that is achieved by distinguishing the current surface of step structures for each position. In this paper, we propose a binarization method based on edge localization to achieve surface distinguishment through the analysis of the contrast and intensity of the coherence signals. After determining the correct envelope, we construct a new form of wavelet family and employ the Morlet wavelet transform to calculate a split point for dividing the anomalous coherence signals. Finally, the effective coherence signals corresponding to the correct envelope are extracted. Surface topography measurement of the nanopillars on the metasurface is obtained from the extracted signals. These nanopillars are distributed with a varying duty cycle and their diameters range from 590 to 1350 nm, near the lateral resolution limit of 501.9 nm of the system. The topography of nanopillars is recovered from the field of view, while the diameter and height of different types of nanopillars are measured. The mean value and standard deviation of these two parameters are consistent with the results obtained via scanning electron microscopy.

Effect of Direct Statistical Contrast Enhancement Technique on Document Image Binarization

Background: Contrast enhancement plays an important role in the image processing field. Contrast correction has performed an adjustment on the darkness or brightness of the input image and increases the quality of the image. Objective: This paper proposed a novel method based on statistical data from the local mean and local standard deviation. Method: The proposed method modifies the mean and standard deviation of a neighbourhood at each pixel and divides it into three categories: background, foreground, and problematic (contrast & luminosity) region. Experimental results from both visual and objective aspects show that the proposed method can normalize the contrast variation problem effectively compared to Histogram Equalization (HE), Difference of Gaussian (DoG), and Butterworth Homomorphic Filtering (BHF). Seven (7) types of binarization methods were tested on the corrected image and produced a positive and impressive result. Result: Finally, a comparison in terms of Signal Noise Ratio (SNR), Misclassification Error (ME), F-measure, Peak Signal Noise Ratio (PSNR), Misclassification Penalty Metric (MPM), and Accuracy was calculated. Each binarization method shows an incremented result after applying it onto the corrected image compared to the original image. The SNR result of our proposed image is 9.350 higher than the three (3) other methods. The average increment after five (5) types of evaluation are: (Otsu = 41.64%, Local Adaptive = 7.05%, Niblack = 30.28%, Bernsen = 25%, Bradley = 3.54%, Nick = 1.59%, Gradient-Based = 14.6%). Conclusion: The results presented in this paper effectively solve the contrast problem and finally produce better quality images.

QBB: Quantile-Based Binarization of 3D Point Cloud Descriptors

Local 3D point feature descriptors play an important role in many areas of computer vision, such as object recognition, registration, etc. There are many well-functioning feature descriptors, but they are typically real-valued and multidimensional vectors, leading to high computational complexity in nearest neighbor searches. To overcome this challenge, methods binarizing real-valued descriptors have emerged. In this paper, we first investigate the available binarization methods and standalone binary feature descriptors and show that existing binarization techniques cannot generally achieve good performance for arbitrary feature descriptors. To remedy this problem, we propose a new binarization method called quantile-based binarization (QBB) that can be applied to any real-valued feature descriptors. It analyses the distribution of feature descriptors that is then used to form meaningful groups along each dimension. To this end, QBB computes quantiles of the empirical distribution and the interval lengths (bin sizes) defined by quantile boundaries. Finally, it assigns a binary code to each group and concatenate them to get the final binary descriptor. QBB is able to adaptively compute the number of bits based on a capacity constraint, i.e., with the appropriate capacity setting, the resulting binary descriptor can be used on devices with lower computational power. We evaluate the descriptiveness of well-known descriptors binarized by QBB and compare them to state-of-the-art methods. According to our evaluation, QBB is able to create binary descriptors whose descriptiveness is closer to the real-valued descriptors than prior approaches. Finally, we also show that QBB can even compete with standalone binary feature descriptors.

A Novel Local Adaptive Percentage Split Distribution Method for Image Binarization and Classification

A Novel Local Adaptive Percentage Split Distribution Method for Image Binarization and Classification

Differential evolutionary topology optimization using shape representation based on KL expansion

Topology optimization is a structural optimization method in which design variables are updated based on the design sensitivity of the objective function and constraints to approach the optimal structure. However, the solution obtained by optimization based on design sensitivity is a local optimum, which is not suitable for nonlinear and multimodal problems. In this study, we propose a topology optimization method based on differential evolution, which is an excellent computational method for global search. While differential evolution has high global search capability, it is computationally expensive, so it is effective to reduce the number of design variables. Therefore, the Karhunen-Loève (KL) expansion is applied to the density function and the terms with small eigenvalues are truncated to express the structural density with a limited number of design variables. We have also proposed a method of binarization by gradually increasing the penalty parameter. In order to construct the optimization algorithm, the optimization is performed for the stiffness maximization problem of a cantilever beam, and the validity of the proposed method is verified by comparing it with the optimization results based on the design sensitivity.

Binary Fruit Fly Swarm Algorithms for the Set Covering Problem

Currently, the industry is experiencing an exponential increase in dealing with binary-based combinatorial problems. In this sense, metaheuristics have been a common trend in the field in order to design approaches to solve them successfully. Thus, a well-known strategy consists in the use of algorithms based on discrete swarms transformed to perform in binary environments. Following the No Free Lunch theorem, we are interested in testing the performance of the Fruit Fly Algorithm, this is a bio-inspired metaheuristic for deducing global optimization in continuous spaces, based on the foraging behavior of the fruit fly, which usually has much better sensory perception of smell and vision than any other species. On the other hand, the Set Coverage Problem is a well-known NP-hard problem with many practical applications, including production line balancing, utility installation, and crew scheduling in railroad and mass transit companies. In this paper, we propose different binarization methods for the Fruit Fly Algorithm, using S-shaped and V-shaped transfer functions and various discretization methods to make the algorithm work in a binary search space. We are motivated with this approach, because in this way we can deliver to future researchers interested in this area, a way to be able to work with continuous metaheuristics in binary domains. This new approach was tested on benchmark instances of the Set Coverage Problem and the computational results show that the proposed algorithm is robust enough to produce good results with low computational cost.

Effective document image binarization via a convex variational level set model

Document image binarization is a significant stage in the optical character recognition system. Different from previous binarization approaches, in this paper, we propose a novel and convex variational level set model for document image binarization. Our energy functional is comprised of two terms: data term and fidelity term. We prove that our model is strictly convex and has a unique global minimum solution, which enables us to set the evolution termination criterion by the normalized step difference energy that measures the convergence state of level set function. We experimentally demonstrate the advantage of the fidelity term and show the merit of level set initialization robustness. In addition, the convergence of the alternative minimization algorithm for solving our model is analyzed. Extensive experiments are conducted on JM dataset, representative degraded document images and DIBCO series datasets to evaluate our model qualitatively and quantitatively. The experiment results verify that our model is effective to deal with most kinds of degraded images. Compared with four evolution and six non-evolution based binarization methods, our model could achieve better or competitive performance in terms of four metrics.

Investigation of the pathophysiology of the retina and choroid in Parkinson's disease by optical coherence tomography

PurposeThe pathology of Parkinson's disease (PD) is suspected to affect the retina and choroid. We investigated changes in the retina and choroid of patients with PD using optical coherence tomography.MethodsWe examined 14 patients with PD and 22 patients without PD. Patients without PD had no ophthalmic disease other than cataracts. In addition, it was also confirmed that there was no neurodegenerative disease. The retinal nerve fiber layer, ganglion cell layer + inner plexiform layer, and choroidal thickness were compared between both groups. Additionally, the choroidal image was divided into the choroid area, luminal area, and interstitial area using the binarization method, and the area of each region and the percentage of luminal area in the choroid area were analyzed.ResultsPatients with PD had a significantly thinner ganglion cell layer + inner plexiform layer compared to those without PD. The choroid area, luminal area, and interstitial area were significantly decreased in patients with PD compared to those without PD. Seven patients with PD who were successfully followed up showed decreased retinal nerve fiber layer and interstitial area after 3 years.ConclusionAutonomic nervous disorders and neurodegeneration in PD can cause thinning of the retina and choroid, as well as a reduction in the choroid area.

Research on Classification Model of Panax notoginseng Taproots Based on Machine Vision Feature Fusion.

The existing classification methods for Panax notoginseng taproots suffer from low accuracy, low efficiency, and poor stability. In this study, a classification model based on image feature fusion is established for Panax notoginseng taproots. The images of Panax notoginseng taproots collected in the experiment are preprocessed by Gaussian filtering, binarization, and morphological methods. Then, a total of 40 features are extracted, including size and shape features, HSV and RGB color features, and texture features. Through BP neural network, extreme learning machine (ELM), and support vector machine (SVM) models, the importance of color, texture, and fusion features for the classification of the main roots of Panax notoginseng is verified. Among the three models, the SVM model performs the best, achieving an accuracy of 92.037% on the prediction set. Next, iterative retaining information variables (IRIVs), variable iterative space shrinkage approach (VISSA), and stepwise regression analysis (SRA) are used to reduce the dimension of all the features. Finally, a traditional machine learning SVM model based on feature selection and a deep learning model based on semantic segmentation are established. With the model size of only 125 kb and the training time of 3.4 s, the IRIV-SVM model achieves an accuracy of 95.370% on the test set, so IRIV-SVM is selected as the main root classification model for Panax notoginseng. After being optimized by the gray wolf optimizer, the IRIV-GWO-SVM model achieves the highest classification accuracy of 98.704% on the test set. The study results of this paper provide a basis for developing online classification methods of Panax notoginseng with different grades in actual production.

Accurate Landmark Localization for Medical Images Using Perturbations

Recently, various studies have been proposed to learn the rich representations of images during deep learning. In particular, the perturbation method is a simple way to learn rich representations that has shown significant success. In this study, we present effective perturbation approaches for medical landmark localization. To this end, we report an extensive experiment that uses the perturbation methods of erasing, smoothing, binarization, and edge detection. The hand X-ray dataset and the ISBI 2015 Cephalometric dataset are used to evaluate the perturbation effect. The experimental results show that the perturbation method forces the network to extract richer representations of an image, leading to performance increases. Moreover, in comparison with the existing methods that lack any complex algorithmic change of network, our methods with specific perturbation methods achieve superior performance.

Method of the optical recognition of technical documentation and the transformation of graphic information into machine-readable form for cognitive analysis

The paper proposes the implementation of the method of optical recognition of technical documentation and the transformation of graphic information into a machine-readable form available for cognitive analysis, which is based on the methods of binarization and alignment of images, text segmentation and recognition. The use of the proposed method will provide a dramatic reduction in the costs of cataloging, checking the completeness and inventory of documentation, as well as an increase in design quality due to the semantic analysis of documentation using a knowledge base that is updated automatically. The article presents the development of the algorithm for optical recognition of a document, preparation of an image for optical recognition of a document, an example of the application of the Sauvola method for binarization of an image, and an analysis of the research results. The proposed implementation allows the text recognition on scanned/photographed documents.

Design of Massive Image Duplicate Search Algorithm in Cloud Computing Environment

Massive image replay algorithm adopts cloud computing technology to realize image uploading, processing and classification on cloud platform. Image classification is realized by image classifier. Zernike matrix was used to compare the classified images and determine the pseudo-edge blocks around the image. According to the pixel value of the image, the binarization method was used to process the image. After processing, edge detection and line detection were carried out to determine the real image size. The similarity of two real images is calculated by normalized similarity criterion, and the image repetition is judged accordingly. The results show that the accuracy of the algorithm is higher than 99%, and the accuracy of the algorithm is higher than that of the comparison algorithm, and the accuracy of the algorithm is the highest when the similarity threshold is 0.8∽0.9.

Integrating Vision System to a Pick and Place Cartesian Robot

Vision aided pick and place cartesian robot is a combination of machine vision system and robotic system. They communicate with each other simultaneously to perform object sorting. In this project, machine vision algorithm for object sorting to solve the problem in failure sorting due to imperfection of images edges and different types of colours is proposed. The image is acquired by a camera and followed by image calibration. Pre-processing of image is performed through these methods, which are HSI colour space transformation, Gaussian filter for image filtering, Otsu’s method for image binarization, and Canny edge detection. LabVIEW edge-based geometric matching is selected for template matching. After the vision application analysed the image, electrical signal will send to robotic arm for object sorting if the acquired image is matched with template image. The proposed machine vision algorithm has yielded an accurate template matching score from 800 to 1000 under different disturbances and conditions. This machine vision algorithm provides more customizable parameters for each methods yet improves the accuracy of template matching.

Microorganism image classification with circle-based Multi-Region Binarization and mutual-information-based feature selection

BackgroundAn important part of water quality control is waterborne pathogen detection. Conventional microorganism detections are done by costly and complicated biochemical approaches, which leave chemical pollutions and require professional laboratories, technicians, and much time. Recently, machine-learning-based image classification algorithms are developed for faster, cheaper, simpler, and more environmentally friendly microorganism detection. Most of them are designed for conventional microscopic images. On the other hand, the optofluidic system is a smaller and cheaper alternative to conventional microscopes, so we aim to propose a microorganism classification algorithm for optofluidic images. MethodIn our previous study, the Binarized-Greyscale-Hybrid Algorithm with Multi-Region Binarization (BiGHAM) algorithm was developed for our optofluidic image dataset. In this paper, a new circle-based region partition method for the Multi-Region Binarization and a new mutual-information-based feature selection method were proposed to improve the BiGHAM method. Additionally, an implementation example of how to adapt and apply our method on an open dataset was presented. Results and conclusionWith the circle-based region partition method and the new feature combination selected based on mutual information, the improved-BiGHAM algorithm achieved higher classification accuracies, and the algorithm is simpler than the previous BiGHAM method during training. For our optofluidic image dataset, the accuracy improvement for the 1000-image-per-class case was from 94.8 to 96.3%, and for the 30-image-per-class case was from 77.8 to 85.6% with the support vector machine as the classifier. The adapted algorithm also works well on the open dataset, showing the potential to be extended to applications on different microorganism images.

Long-term effect of intravitreal aflibercept treatment on choroidal vascularity index in neovascular age-related macular degeneration

BackgroundOur aim is to examine the effect of intravitreal aflibercept treatment on the choroidal vascularity index in eyes with neovascular age-related macular degeneration (nARMD) using the binarization method. MethodsIncluded in the study were the 38 eyes of 38 patients diagnosed with nARMD and treated with intravitreal aflibercept at our clinic. The patients’ 12-month records were reviewed retrospectively. Patients’ choroidal thickness (CT) was measured at baseline, 3rd month, 6th month, and 12th month on EDI-OCT images. We calculated luminal area (LA), stromal area (SA), total choroidal area (TCA), and choroid vascular index (CVI) values using binarization. ResultsThe mean age of the patients was 70,1 ± 8 years. The mean of injections over 12 months was 5.6 ± 2. Baseline values were CT 224 µm, LA 0,214 mm2, SA 0,119mm2, TCA 0,333 mm2, and CVI 64% respectively. At the 3rd month, values were CT 223 µm, LA 0,212 mm2, SA 0,121 mm2, TCA 0,332 mm2, and CVI 64%, respectively. The 6th-month values were CT 221 µm, LA 0,204 mm2, SA 0,116 mm2, TCA 0,320 mm2, and CVI 64%, respectively, and 12-month values were CT 218 µm, LA 0,204 mm2, SA 0,112 mm2, TCA 0,318 mm2, and CVI 64%, respectively. According to LogMAR, BCVA was 0.59 at baseline, 0.56 at 3 months, 0.52 at 6 months, and 0.51 at 12 months. At the end of month 12, we observed a statistically significant decrease in CT. No significant changes were observed in other parameters. When the number of injections was correlated with CVI, CT, and BCVA, a significant negative correlation was observed with CVI ConclusionsAflibercept treatment reduces CT by affecting not only luminal but also stromal areas of the choroid in the long term but does not change the CVI. However, as the number of injections increases, CVI decreases.