Pixel Precision Research Articles

A Nonintrusive Optical Approach to Characterize Heliostats in Utility-Scale Power Tower Plants: Camera Position Sensitivity Analysis

Abstract Optics plays a major role in the effectiveness of concentrating solar power (CSP) technologies. The nonintrusive optical (NIO) approach developed by the National Renewable Energy Laboratory uses uncrewed aircraft system (UAS)-based imaging to survey heliostats in a commercial-scale power tower CSP plant and characterize their optical errors. The image processing algorithm uses photogrammetry to calculate the camera position for each image frame, and the accuracy of the estimated optical errors is highly sensitive to the calculated camera position accuracy. In this study, we simulate a series of case studies in python to examine the impact of different parameters of the sensitivity of the camera calculation, including the number of facet corners used as control points for the photogrammetric calculation, precision error in the detected pixel locations of the facet corners in the image, and precision error of the canting and mounting positions of the facets of the heliostat. The case studies consider heliostat geometry based on three commercial designs to serve as representative examples of different possible sizes of heliostats that the NIO method could be applied to. The results show that increasing the number of control points can improve accuracy for heliostats with many facets, pixel precision has a significantly larger impact on camera calculation accuracy than facet canting and mounting errors, and the camera distance and focal length must be chosen to ensure adequate pixel accuracy on the heliostat surface depending on the size of heliostat. Based on the results, recommendations for the allowable values of each parameter are provided to achieve the required NIO optical error estimation accuracy depending on the size of heliostat.

Remote sensing image road network detection based on channel attention mechanism

Extracting and detecting road network consistency from high-resolution remote sensing images has been a hot and difficult problem in the computer vision. Although it has made significant progress, there is still a phenomenon of high training accuracy but unsatisfactory actual extraction and detection results. The attention mechanism is one of the efficient and practical mechanisms in deep learning. It improves the performance of deep learning by selectively focusing on a portion of all information while ignoring other visible information, while effectively utilizing computing resources. Numerous experiments have also confirmed that the attention mechanism is resource-saving and effective. Its plug and play feature brings great convenience to programmers. In order to provide better road network detection results and solve the above problem, this paper combines the channel attention mechanism with ResNet and proposes SE-ResNet and ECA-ResNet for remote sensing image road network detection, making networks extract and learn road network features and ignore some non-road network features. The experimental results show that on the Massachusetts roads (MR) and CHN6-CUG roads datasets, ECA-ResNet and SE-ResNet based on channel attention mechanism perform similar to LeNet7 and ResNet in terms of accuracy, loss, accuracy convergence, and loss convergence, and even increase a certain computational burden. However, their final road network detection results (including road network detection pixel count, precision, recall, accuracy, IOU, F1 score, and actual road network detection result) of the former are significantly better than those of the latter. The channel attention mechanism makes the deep neural network pay more attention to the extraction and learning of road network features, while ignoring the extraction and learning of some non-road network features, which improves the accuracy of containing road network samples and reduces the accuracy of not containing road network samples. Therefore, the performance of ECA-ResNet and SE-ResNet is similar to that of LeNet7 and ResNet in the accuracy, loss, accuracy convergence and loss convergence, but the final road network detection results of ECA-ResNet and SE-ResNet are significantly better than those of LeNet7 and ResNet. Therefore, the proposed ECA-ResNet and SE-ResNet have broad application prospects in road network detection, especially ECA-ResNet.

Real Time Object Detection in Low Light Environment Using Adaptive Image Information Enhancement Algorithm

Object detection plays a vital role in CCTV systems. Cameras are now deployed at traffic lights, roadways, shopping centers, railways, banks, and other public locations to increase security. It isn't easy to follow the video quickly and continuously, though. Therefore, surveillance cameras are not required, and human surveillance is needed. A significant difficulty for CCTV cameras is detecting anomalies such as theft, accidents, crimes, and other illegal activities. The frequency of abnormal behaviour is the same as that of regular events. To detect objects in a video, we first analyze each pixel in the image. Segmentation in digital photography is the process of dividing different parts of an image into pixels. Segmentation performance is affected by uneven and dim lighting. These factors significantly impact the realtime object detection process of CCTV systems. In this article, we propose an Adaptive Image Information Enhancement Algorithm (AIIE) to improve images affected by poor lighting. Test results compare the output of the current method with the improved ResNet model architecture and show significant improvement in object detection in video streams. The proposed model delivers better results regarding metrics such as precision, recall, and pixel precision. We also saw significant improvements in object detection.

An analysis of precision: occlusion and perspective geometry’s role in 6D pose estimation

Achieving precise 6 degrees of freedom (6D) pose estimation of rigid objects from color images is a critical challenge with wide-ranging applications in robotics and close-contact aircraft operations. This study investigates key techniques in the application of YOLOv5 object detection convolutional neural network (CNN) for 6D pose localization of aircraft using only color imagery. Traditional object detection labeling methods suffer from inaccuracies due to perspective geometry and being limited to visible key points. This research demonstrates that with precise labeling, a CNN can predict object features with near-pixel accuracy, effectively learning the distinct appearance of the object due to perspective distortion with a pinhole camera. Additionally, we highlight the crucial role of knowledge about occluded features. Training the CNN with such knowledge slightly reduces pixel precision, but enables the prediction of 3 times more features, including those that are not initially visible, resulting in an overall better performing 6D system. Notably, we reveal that the data augmentation technique of scale can interfere with pixel precision when used during training. These findings are crucial for the entire system, which leverages the Solve Perspective-N-Point (Solve-PnP) algorithm, achieving 6D pose accuracy within 1^circ and 7 cm at distances ranging from 7.5 to 35 m from the camera. Moreover, this solution operates in real-time, achieving sub-10ms processing times on a desktop PC.

DME-DeepLabV3+: a lightweight model for diabetic macular edema extraction based on DeepLabV3+ architecture.

Diabetic macular edema (DME) is a major cause of vision impairment in the patients with diabetes. Optical Coherence Tomography (OCT) is an important ophthalmic imaging method, which can enable early detection of DME. However, it is difficult to achieve high-efficiency and high-precision extraction of DME in OCT images because the sources of OCT images are diverse and the quality of OCT images is not stable. Thus, it is still required to design a model to improve the accuracy of DME extraction in OCT images. A lightweight model (DME-DeepLabV3+) was proposed for DME extraction using a DeepLabV3+ architecture. In this model, MobileNetV2 model was used as the backbone for extracting low-level features of DME. The improved ASPP with sawtooth wave-like dilation rate was used for extracting high-level features of DME. Then, the decoder was used to fuse and refine low-level and high-level features of DME. Finally, 1711 OCT images were collected from the Kermany dataset and the Affiliated Eye Hospital. 1369, 171, and 171 OCT images were randomly selected for training, validation, and testing, respectively. In ablation experiment, the proposed DME-DeepLabV3+ model was compared against DeepLabV3+ model with different setting to evaluate the effects of MobileNetV2 and improved ASPP on DME extraction. DME-DeepLabV3+ had better extraction performance, especially in small-scale macular edema regions. The extraction results of DME-DeepLabV3+ were close to ground truth. In comparative experiment, the proposed DME-DeepLabV3+ model was compared against other models, including FCN, UNet, PSPNet, ICNet, and DANet, to evaluate DME extraction performance. DME-DeepLabV3+ model had better DME extraction performance than other models as shown by greater pixel accuracy (PA), mean pixel accuracy (MPA), precision (Pre), recall (Re), F1-score (F1), and mean Intersection over Union (MIoU), which were 98.71%, 95.23%, 91.19%, 91.12%, 91.15%, and 91.18%, respectively. DME-DeepLabV3+ model is suitable for DME extraction in OCT images and can assist the ophthalmologists in the management of ocular diseases.

RGB image-based method for phenotyping rust disease progress in pea leaves using R

BackgroundRust is a damaging disease affecting vital crops, including pea, and identifying highly resistant genotypes remains a challenge. Accurate measurement of infection levels in large germplasm collections is crucial for finding new resistance sources. Current evaluation methods rely on visual estimation of disease severity and infection type under field or controlled conditions. While they identify some resistance sources, they are error-prone and time-consuming. An image analysis system proves useful, providing an easy-to-use and affordable way to quickly count and measure rust-induced pustules on pea samples. This study aimed to develop an automated image analysis pipeline for accurately calculating rust disease progression parameters under controlled conditions, ensuring reliable data collection.ResultsA highly efficient and automatic image-based method for assessing rust disease in pea leaves was developed using R. The method’s optimization and validation involved testing different segmentation indices and image resolutions on 600 pea leaflets with rust symptoms. The approach allows automatic estimation of parameters like pustule number, pustule size, leaf area, and percentage of pustule coverage. It reconstructs time series data for each leaf and integrates daily estimates into disease progression parameters, including latency period and area under the disease progression curve. Significant variation in disease responses was observed between genotypes using both visual ratings and image-based analysis. Among assessed segmentation indices, the Normalized Green Red Difference Index (NGRDI) proved fastest, analysing 600 leaflets at 60% resolution in 62 s with parallel processing. Lin’s concordance correlation coefficient between image-based and visual pustule counting showed over 0.98 accuracy at full resolution. While lower resolution slightly reduced accuracy, differences were statistically insignificant for most disease progression parameters, significantly reducing processing time and storage space. NGRDI was optimal at all time points, providing highly accurate estimations with minimal accumulated error.ConclusionsA new image-based method for monitoring pea rust disease in detached leaves, using RGB spectral indices segmentation and pixel value thresholding, improves resolution and precision. It rapidly analyses hundreds of images with accuracy comparable to visual methods and higher than other image-based approaches. This method evaluates rust progression in pea, eliminating rater-induced errors from traditional methods. Implementing this approach to evaluate large germplasm collections will improve our understanding of plant-pathogen interactions and aid future breeding for novel pea cultivars with increased rust resistance.

P‐75: Development of Medium and Large Size Active Bistable State Dual&Four Color Cholesteric LCDs †

Compared with cholesteric display technology, which is mainly a passive driver under 5inch, this paper report one 7.5inch actively driven bistable display device. We use cyan cholesteric liquid crystal reflecting 528nm light and red ink to achieve red & white dual color display. In order to enrich the display information, one red (660nm) cell and one cyan (528nm) cell are fitted with pixel precision, and black ink is printed on the bottom, control the status of the two cells respectively, so as to achieve red, cyan, black, and white four color display. In addition, we optimized drive mode, and update traditional two‐stage drive to three‐stage drive, which improved the stability of the LCD FC state and the image quality of the display device, especially at high and low temperatures. In this paper, red&white dual color and four color display technologies are the first in the LCD industry.

Detection and counting of banana bunches by integrating deep learning and classic image-processing algorithms

Robots must first detect the number of banana bunches when making judgements on sterile bud removal and estimating weight for harvest in the field environment. Banana bunches are complex in shape, arranged in a nonlinear helical curve along the stalk, and have different growth states in different periods, with bunches widely spaced in the early period and densely arranged in the harvest period. Deep learning nor classical image-processing algorithms alone can detect and count bunches in both periods. Therefore, these algorithms were combined to calculate the number of bunches in the two periods. For counting bunches in the debudding period, the convolutional neural network Deeplab V3 + model and classic image-processing algorithm were combined to finely segment bunches and calculate bunch numbers, providing intelligent decision-making for judgment on the timing for debudding. To count bunches during harvest, based on deep learning to identify the overall banana fruit cluster, the edge detection algorithm was employed to extract the centroid points of fruit fingers, and the clustering algorithm was used to determine the optimal number of bunches on the visual detection surface. An estimation model for the total number of bunches, including hidden ones, was created based on their helical curve arrangement. The results indicated a target segmentation MIoU of 0.878 during the debudding period, a mean pixel precision of 0.936, and a final bunch detection accuracy rate of 86%. Bunch detection was highly challenging during the harvest period, with a detection accuracy rate of 76% and a final overall bunch counting accuracy rate of 93.2%. Software was designed to estimate banana fruit weight during the harvest period. This research method provided a theoretical basis and experimental data support for automatic sterile bud removal and weight estimation for bananas.

AFU-Net: A Novel U-Net Network for Rice Leaf Disease Segmentation

Highlights The attention mechanism enhances the ability of the model to learn specific semantic information in encoder. The redesigned residual structure deepens the network while reducing the number of parameters. The feature extraction module and feature fusion module obtain richer boundary feature information and effectively integrate output results from different levels. The mIoU, mPA, and Precision values of AFU-Net in the self-built dataset are 87.25%, 92.23%, and 99.67%, respectively. Abstract. Rice diseases adversely affect rice growth and yield. Precise spot segmentation helps to assess the severity of the disease so that appropriate control measures can be taken. In this article, we propose a segmentation method called AFU-Net for rice leaf diseases, and its performance is verified through experiments. Based on the traditional UNet, this method incorporates an attention mechanism, a residual module and a feature fusion module (FFM). The attention mechanism is embedded in skip connections, which enhances the learning of particular semantic features in the encoder layer. In addition, the residual module is integrated into the decoder layer, which deepens the network and enables it to extract richer semantic information. The proposed FFM structure effectively enhances the learning of boundary information and local detail features. The experimental results show that the mean intersection over union (mIoU), mean pixel accuracy (mPA) and Precision of the proposed model on the self-built rice leaf disease segmentation dataset are 87.25%, 92.23%, and 99.67%, respectively. All three evaluation indexes were improved over the control group, while the proposed model had the lowest number of parameters and displayed a good segmentation effect for smaller disease points and disease parts with less obvious characteristics. Keywords: Attention mechanism, Feature fusion module, Residual module, Rice leaves, UNet model.

A high-precision Mark positioning algorithm based on sub-pixel shape template matching in wafer bonding alignment

Mark's precise positioning is critical for achieving wafer bonding accuracy. A Mark positioning algorithm is proposed to improve the accuracy and stability of Mark positioning. Shape template matching can effectively reduce the impact of image quality on positioning accuracy, but it can only achieve pixel precision Mark location. On the basis of this method, the algorithm uses the facet model method to extract the sub-pixel edge information for the Mark contour, and combines the sub-pixel traversal strategy to achieve accurate sub-pixel precision Mark location. Simultaneously, the composite template method is proposed for the first time to improve the template information's quality, which effectively improves the robustness of the algorithm to image quality. The simulation results indicate that the algorithm is capable of positioning with an accuracy of approximately 0.04 pixels. The experimental results indicate that the algorithm is capable of positioning with an accuracy of approximately 0.052 pixels and is resistant to image noise and blur, which meets the accuracy and stability requirements for Mark positioning in wafer bonding alignment.

Non-Newtonian fluid simulation and reconstruction from monocular videos

In fluid simulation, the parameters of the constitutive model are difficult to predict accurately, which leads to the inconsistency between the simulation results and the visual effects of monocular videos. To solve this problem, a monocular video-guided non-Newtonian fluid reconstruction method is proposed. The model takes non-Newtonian fluid simulation videos as input for training and learns the best low-dimensional latent space representation of a single frame of fluid simulation images. Inter-frame prediction is then performed in this latent domain, employing a convolutional long–short-term memory network to predict latent vector representations of future frames. Finally, the reconstruction parameters of the constitutive model are predicted based on the frame-by-frame latent representation encoding and inter-frame temporal features. In the model verification stage, the real video of the non-Newtonian fluid is used as the input to predict the parameters of the fluid constitutive model, and realize the simulation and reconstruction of the non-Newtonian fluid based on the Cross model. Experimental results showed that the video-guided simulation reconstruction method could obtain fluid flow phenomena that were more consistent with those in the real video than the reconstruction method based on rheometer measurements. Furthermore, it provided higher pixel accuracy and precision at different times and visual effects that were more consistent with the actual fluid flow.

Comparative studies of deep learning segmentation models for left ventricle segmentation.

One of the primary factors contributing to death across all age groups is cardiovascular disease. In the analysis of heart function, analyzing the left ventricle (LV) from 2D echocardiographic images is a common medical procedure for heart patients. Consistent and accurate segmentation of the LV exerts significant impact on the understanding of the normal anatomy of the heart, as well as the ability to distinguish the aberrant or diseased structure of the heart. Therefore, LV segmentation is an important and critical task in medical practice, and automated LV segmentation is a pressing need. The deep learning models have been utilized in research for automatic LV segmentation. In this work, three cutting-edge convolutional neural network architectures (SegNet, Fully Convolutional Network, and Mask R-CNN) are designed and implemented to segment the LV. In addition, an echocardiography image dataset is generated, and the amount of training data is gradually increased to measure segmentation performance using evaluation metrics. The pixel's accuracy, precision, recall, specificity, Jaccard index, and dice similarity coefficients are applied to evaluate the three models. The Mask R-CNN model outperformed the other two models in these evaluation metrics. As a result, the Mask R-CNN model is used in this study to examine the effect of training data. For 4,000 images, the network achieved 92.21% DSC value, 85.55% Jaccard index, 98.76% mean accuracy, 96.81% recall, 93.15% precision, and 96.58% specificity value. Relatively, the Mask R-CNN outperformed other architectures, and the performance achieves stability when the model is trained using more than 4,000 training images.

Surface topography data fusion of additive manufacturing based on confocal and focus variation microscopy.

In this paper, two innovative data fusion methods are proposed for reconstructing the surfaces produced by directed energy deposition (DED) additive manufacturing. The surface topographic data were obtained from confocal laser scanning microscopy (CLSM) and focus variation microscopy (FV). The first method (competitive data fusion) aims to improve the data quality by combining both the advantages of the CLSM and FV techniques, while the second method (cooperative data integration) is designed for generating a single representation that contains not only global information but also local details. The results show that both fusion methods achieved satisfactory results: in the competitive fusion, the fused data preserved the characteristics of FV data while its vertical resolution is also improved by integrating the short waves from the CLSM data; the cooperative data fusion achieved one pixel precision of the surface registration which adopted the feature-based registration method with the help of color image information. The computational complexity is reduced from O((m×n)2) to O(m×n + k). Both proposed data fusion methods provided innovative solutions for the microscopic surface reconstruction and surface representation in multiscales in the field of additive manufacturing.

A machine learning model for separating epithelial and stromal regions in oral cavity squamous cell carcinomas using H&E-stained histology images: A multi-center, retrospective study

ObjectiveTissue slides from Oral cavity squamous cell carcinoma (OC-SCC), particularly the epithelial regions, hold morphologic features that are both diagnostic and prognostic. Yet, previously developed approaches for automated epithelium segmentation in OC-SCC have not been independently tested in a multi-center setting. In this study, we aimed to investigate the effectiveness and applicability of a convolutional neural network (CNN) model to perform epithelial segmentation using digitized H&E-stained diagnostic slides from OC-SCC patients in a multi-center setting. MethodsA CNN model was developed to segment the epithelial regions of digitized slides (n = 810), retrospectively collected from five different centers. Deep learning models were trained and validated using well-annotated tissue microarray (TMA) images (n = 212) at various magnifications. The best performing model was locked down and used for independent testing with a total of 478 whole-slide images (WSIs). Manually annotated epithelial regions were used as the reference standard for evaluation. We also compared the model generated results with IHC-stained epithelium (n = 120) as the reference. ResultsThe locked-down CNN model trained on the TMA image training cohorts with 10x magnification achieved the best segmentation performance. The locked-down model performed consistently and yielded Pixel Accuracy, Recall Rate, Precision Rate, and Dice Coefficient that ranged from 95.8% to 96.6%, 79.1% to 93.8%, 85.7% to 89.3%, and 82.3% to 89.0%, respectively for the three independent testing WSI cohorts. ConclusionThe automated model achieved a consistently accurate performance for automated epithelial region segmentation compared to manual annotations. This model could be integrated into a computer-aided diagnosis or prognosis system.

Coastal Monitoring Using Unmanned Aerial Vehicles (UAVs) for the Management of the Spanish Mediterranean Coast: The Case of Almenara-Sagunto.

The concentration of the world’s population in coastal areas means an increase in pressure on the environment and coastal ecosystems. The impacts of climate change affect natural biophysical and ecological systems and human health. Research has been developed to create coastal monitoring with Unmanned Aerial Vehicles (UAVs) that allow data to be obtained and methodologies that integrate computer vision algorithms for 3D and image processing techniques for analysis, combined with maritime information. The Valencian oval is located on the Spanish Mediterranean coast and registers significant coastal erosion. It is a densely populated area, with high economic relevance and tourist activity. The main goals of the developed research in this coastal area include creating a methodology of data collection that identifies environmental indicators significant to community health and uses in the coastal areas, to test progression of interventions and to assess coastal erosion detection and monitoring. The final objective is to aid in decision-making and coastal management. Sediment characterization was obtained, and continuous maritime information was collected. The dynamic evolution of coastal areas was researched by using UAVs on the Spanish Mediterranean coast. This technique is suitable for measuring medium to small coastal changes. Flight planning was carried out using the grid mode and adapted to areas in order to obtain a homogeneous pixel size and precision. This monitoring program takes advantage of technological development with very low economic costs and is a good tool for making decisions that must be based on scientific information. With the monitoring work, an annual erosion between 12 and 6 m was detected. The monitoring program has evidenced the shoreline trend as a result of the impact of rigid structures, mainly ports and groins, in promoting down-drift erosion processes in the area.

Land Use Classification Using Improved U-Net in Remote Sensing Images of Urban and Rural Planning Monitoring

Aiming at the problem of low accuracy and efficiency of existing land use classification methods for high-resolution remote sensing image segmentation, a land use classification method using improved U-Net in remote sensing images of urban and rural planning monitoring is proposed. First, taking the high-resolution remote sensing images of different remote sensing satellites as the data source, the remote sensing images in the data source are registered and cropped so that the pixels at the corresponding positions represent the same geographical location. Then, the encoder of the U-Net model is combined with the residual module to share the network parameters and avoid the degradation of the deep network. The dense connection module is integrated into the decoder to connect the shallow features with the deep features, so as to obtain new features and improve the feature utilization rate. Finally, the depthwise separable convolution is used to process the spatial and channel information of the convolution process separately to reduce the model parameters. Experiments show that the pixel accuracy, recall rate, precision rate, and average intersection-over-union ratio of the proposed land use classification method based on improved U-Net are 92.35%, 80.56%, 83.45%, and 86.75%, respectively, which are better than the compared methods. Therefore, the proposed method is proved to have good land use classification ability.

Tomography analysis tool: an application for image analysis based on unsupervised machine learning

We developed a graphical user interface (GUI) to analyse tomographic images of superconducting Nb3Sn wires designed for the next generation accelerator magnets. The Tomography Analysis Tool (TAT) relies on the k-means algorithm, an unsupervised machine learning technique which is widely used to partition images into separated clusters. The GUI is compatible with both Linux and Windows operating systems. The software reliability was tested by optical inspecting the tomographic images superimposed on the clustered image obtained by the k-means algorithm. TAT was proven to correctly segment the various components of the Nb3Sn superconducting wires with single pixel precision. Finally, this software can be a useful tool for the scientific community to segment and analyse quickly and reproducibly tomographic images.

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.

Segmentation of Squamous Columnar Junction on VIA Images using U-Net Architecture

Cervical cancer is the second most common cancer that affects women, especially in developing countries including Indonesia. Cervical cancer is a type of cancer found in the cervix, precisely in the squamous columnar junction (SCJ). Early screening for cervical cancer can be reduce the risk of cervical cancer. One of the popular screening tool methods for the detection of cervical pre-cancer is the Visual Inspection with Acetic Acid (VIA) method. This is due to the level of effectiveness, convenience and low cost. This paper proposes a method for the detection and segmentation of the SCJ region on VIA images using U-Net. This study is the first research conducted using the CNN method to perform segmentation tasks in the SCJ region. The best performance results are shown from the Pixel Accuracy, Mean IoU, Mean Accuracy, Dice coefficient, Precision and Sensitivity values, namely 90.86%, 56.5%, 75.69%, 34.09%, 41.24%, and 56.91%.
 Keywords: Cervical Pre-cancer, Screening VIA, SCJ, U-Net.

Detecting and analysing geomorphological structures in images of comet 67P/Churyumov–Gerasimenko using Fourier transform

ABSTRACT We present a method for automatized detection and analysis of quasi-periodic lineament structures from images at pixel precision. The method exploits properties of the images’ frequency domain found by using the Fourier transform. We developed this method with the goal of detecting lineament structures in an image of the Hathor cliff of comet 67P/Churyumov–Gerasimenko, which are caused by layerings and furrows in the nucleus material. Using our method, we determined the orientation and wavelength range of these structures. The detected layering edges have similar orientations and spatial separations of 9–20 m, and are ubiquitous throughout the image. We suggest that the layerings are a global feature of the comet nucleus that provide information about formation and evolution of comet 67P. The furrows are non-uniformly distributed throughout the image. Their orientation is broadly parallel to the direction of the local gravity vector at the Hathor cliff, with spacings similar to those of the layering structures. The furrows are interpreted as signatures of local down-slope movement of cliff material. We demonstrate that the developed method is broadly applicable to the detection and analysis of various kinds of quasi-periodic structures like geological layering, folding and faulting, and texture analysis in general. In order to facilitate the application of our method, this paper is accompanied by a demo program written in matlab.