Search

The white blood cell (WBC) plays an instrumental role in protecting the human body from different infectious diseases and unknown biological invaders. Their count and cellular morphology drastically change when the body suffers from any such disease or abnormal conditions. This paper demonstrates the detection of five types of WBC present in human blood smear images. Several authors proposed WBC detection using only one color space. But this paper addresses WBC-type detection using four popular color spaces – RGB (red, green, and blue), HSV (hue, saturation, and intensity), YCbCr (luminance, chroma blue, and chroma red), and L*a*b* (luminosity, chromaticity layer "a*", and chromaticity layer "b*"). The proposed methodology involves color thresholding, K-means clustering, nuclear region analysis, and the implementation of a feedforward neural network. The unsupervised machine learning algorithm, K-means clustering is used to segment nuclei from the blood smear images. The detection of WBC-type using four color spaces is compared in terms of accuracy, sensitivity, specificity, and positive predictive value. It is observed that HSV-based WBC-type detection methodology outperforms methodologies based on RGB, YCbCr, and L*a*b* color spaces. The same methodology mentioned above is then compared with the works proposed by other authors. [ABSTRACT FROM AUTHOR]

This paper presents a novel automatic system to estimate the traffic light signal of interest for a vehicle driver in colored images of multiple lanes street from a camera mounted on car. The system analyzes the image to determine the candidate regions for traffic light signals. These regions are categorized based on their color and shape. Then spatial constraints are applied to generate an estimate of the traffic location. Existing solutions attempt to detect all traffic light signals in an image. Further solutions require prior information to specify the traffic light of interest with GPS information. But there is no work that estimate the traffic light signal of interest for the driver when multiple traffic light signals present in an image without prior information. There are two key contributions for this work: Building the first prototypical system that generates an estimate for the traffic light signal, and using a unique combination of the normalized RGB and the HSV color spaces. The proposed system was evaluated on LISA Traffic Light Dataset with maximum accuracy 97.6%. Moreover, the proposed system has been tested on private data sets from real-world environments. This work helps to improve the performance of advanced driver assistance systems. [ABSTRACT FROM AUTHOR]

Coandă effect screen structures are used to protect hydropower turbines in mountainous regions from impurities of different sizes. To save time during the design phase, computational fluid dynamics simulations (CFD) can support experimental investigations. Experimental investigations and 2D CFD simulations are performed on the base of the Coandă effect screen structure. Flow rates of 7 m3/h, 8 m3/h, and 9 m3/h are simulated, captured by a camera system in the experiments, and compared. In the experimental setting, the fill level of the Coandă rake is measured. For comparison, the RMSE of the y-position fill level is computed between simulation and measurement. Furthermore, the mass flow balance, Π , is evaluated. The RMSE of the y-direction fill level between simulation and measurement is on average 2.1 mm and 4.4 mm for the lower CFD level and the higher CFD level, respectively. Experiments show an average 6.72% higher Π. The reason for this is that 2D CFD does not consider the surface tension of the borders. In the diverted flow region, the flow field based on experiments appears less viscous than in the simulations. CFD simulations are an appropriate tool to support the design process of Coandă effect screens, especially for fill level determination. The validated CFD code can reduce the cost and time of experiments to design future Coandă effect screen structures. [ABSTRACT FROM AUTHOR]

Coandă effect screen structures are used to protect hydropower turbines in mountainous regions from impurities of different sizes. To save time during the design phase, computational fluid dynamics simulations (CFD) can support experimental investigations. Experimental investigations and 2D CFD simulations are performed on the base of the Coandă effect screen structure. Flow rates of 7 m3/h, 8 m3/h, and 9 m3/h are simulated, captured by a camera system in the experiments, and compared. In the experimental setting, the fill level of the Coandă rake is measured. For comparison, the RMSE of the y-position fill level is computed between simulation and measurement. Furthermore, the mass flow balance, Π, is evaluated. The RMSE of the y-direction fill level between simulation and measurement is on average 2.1 mm and 4.4 mm for the lower CFD level and the higher CFD level, respectively. Experiments show an average 6.72% higher Π. The reason for this is that 2D CFD does not consider the surface tension of the borders. In the diverted flow region, the flow field based on experiments appears less viscous than in the simulations. CFD simulations are an appropriate tool to support the design process of Coandă effect screens, especially for fill level determination. The validated CFD code can reduce the cost and time of experiments to design future Coandă effect screen structures.

Low resolution mobile photographed images pose a complex set of research challenges as compared to non-mobile captured images, which really is a significant issue these days. For non-mobile captured and high-resolution photos, current plant recognition systems are the best solution providers. This study proposes the identification and extraction of leaf regions from complex backgrounds to meet the automatic recognition needs of a variety of mobile phone users. Additionally multiple factors complicate the leaf region extraction from complex backgrounds such as varying background patterns, clutters, varying leaf shape/size and varying illumination due to volatile weather conditions. In this paper, a simple and efficient method for leaf extraction from complex background of mobile photographed low resolution images is proposed based on color channel thresholding and morphological operations. A self-built database of 5000 mobile photographed images in realistic environments is adapted for experimentations. Experiments were conducted on various resolution categories, and it was discovered that the proposed model has an average dice similarity measure of 99.5 percent for successful extraction of the leaf region in 13MP mobile photographed images. Furthermore, our comparative investigation reveals that the suggested model outperforms both traditional and state-of-the-art techniques. [ABSTRACT FROM AUTHOR]

In developing countries, the cotton harvesting operation is currently being performed manually. Due to the monotonous nature of this task and the involvement of a considerable amount of labor, this operation becomes very tedious and costly. The harvesting robots can be a good alternative for the selective picking of cotton bolls from the field. In this study, an attempt has been made to develop the image processing algorithms for in-field cotton boll detection in natural lighting conditions for the cotton harvesting robot. Four image processing algorithms namely color difference, band ratio, YCbCr method, and chromatic aberration were proposed for the real-time segmentation of cotton bolls under natural outdoor light conditions. The performance of developed image processing algorithms was evaluated and the experimental results revealed that the chromatic aberration method outperforms as compared to other developed algorithms. The chromatic aberration method showed the highest identification rate of 91.05% with false positive and false negative rates of 6.99% and 4.88% respectively, among all the proposed algorithms. The highest sensitivity and specificity were found to be 81.31% and 97.53%, respectively, using the chromatic aberration method. Overall, the chromatic aberration approach demonstrated a very promising performance for in-field cotton bolls detection under natural lighting conditions which confirms its applicability for the robotic cotton harvesters.

Periodontal (gum) diseases, reportedly affect 4 out of 10 adults 30 years of age or older in the USA. The standard of care for clinical assessments of these diseases is bleeding on probing, which is invasive, subjective and semi-qualitative. Thus, research on proposing alternative noninvasive modalities for clinical assessments of periodontal tissues is crucial. Quantitative Ultrasound (QUS) has shown promises in noninvasive assessments of various diseases in soft biological tissues; however, it has not been employed in periodontology. Here as the first step, we focused on QUS-based characterization of two very adjacent oral soft tissues of alveolar mucosa and attached gingiva in an in vivo animal study. We investigated first order ultrasonic speckle statistics using the two-parameter Burr model (power-law b and scale factor l ). Our QUS analysis was compared with the Masson's Trichrome histology images of the two oral tissue types quantitatively using the RGB color thresholding. QUS study included 10 swine and US scanning was performed at the first and second molars of all four oral quadrants in each swine, resulting in 80 scans. US scan data was acquired at the transit/receive frequency of 24 MHz using a toothbrush-sized transducer. Parametric imaging of Burr parameters was created using a sliding kernel method with linear interpolations. The kernel size and overlap ratio was 10 wavelengths and 70%, respectively. No statistically significant difference was reported for estimated parameters when interpolation was performed ( p-value >0.01). Results at both oral sites (molar 1 and molar 2) showed that the difference between the two tissue types using Burr parameters were statistically significant ( p-value <0.0001). The average Burr b was reported to be higher in attached gingiva while the average Burr l was lower compared to mucosa. Visual comparison of Masson Trichrome histology images of these tissues showed denser color density in gingiva. The color thresholding of these images further confirmed that the percent of blue, which stains collagenous regions, was at least two times higher in gingiva than alveolar mucosa, based on threshold values. Comparing histology and QUS in characterizing the two tissue types, it was suggested that the elevated Burr b (related to potential scatterer densities) in gingiva could be aligned with findings from Masson's Trichrome histology. This study showed a promising potential of QUS for periodontal soft tissue characterization.

Wild-type transthyretin amyloidosis (ATTRwt) is an underdiagnosed and potentially fatal disease. Interestingly, ATTRwt deposits have been found to deposit in the ligamentum flavum (LF) of patients with lumbar spinal stenosis before the development of systemic and cardiac amyloidosis. In order to study this phenomenon and its possible relationship with LF thickening and systemic amyloidosis, a precise method of quantifying amyloid deposits in histological slides of LF is critical. However, such a method is currently unavailable. Here, we present a machine learning quantification method with Trainable Weka Segmentation (TWS) to assess amyloid deposition in histological slides of LF. Images of ligamentum flavum specimens stained with Congo red are obtained from spinal stenosis patients undergoing laminectomies and confirmed to be positive for ATTRwt. Amyloid deposits in these specimens are classified and quantified by TWS through training the algorithm via user-directed annotations on images of LF. TWS can also be automated through exposure to a set of training images with user-directed annotations, and then applied] to a set of new images without additional annotations. Additional methods of color thresholding and manual segmentation are also used on these images for comparison to TWS. We develop the use of TWS in images of LF and demonstrate its potential for automated quantification. TWS is strongly correlated with manual segmentation in the training set of images with user-directed annotations (R = 0.98; p = 0.0033) as well as in the application set of images where TWS was automated (R = 0.94; p = 0.016). Color thresholding was weakly correlated with manual segmentation in the training set of images (R = 0.78; p = 0.12) and in the application set of images (R = 0.65; p = 0.23). TWS machine learning closely correlates with the gold-standard comparator of manual segmentation and outperforms the color thresholding method. This novel machine learning method to quantify amyloid deposition in histological slides of ligamentum flavum is a precise, objective, accessible, high throughput, and powerful tool that will hopefully pave the way towards future research and clinical applications.

Breast cancer is the common existing form of cancers amongst women. The automatic image analysis methods have an enormous potential to decrease the workload in a pathology laboratory. The grading of breast cancer histopathology images is used to find the level of breast cancer. The automatic grading of breast cancer histopathology images is a challenging task. In this paper a system for automatic detection of breast cancer grading of H&E stained histopathological images is presented. An image processing techniques such as preprocessing, segmentation, feature extraction and classification are used in this system. The segmentation of nuclei in H&E stained image is performed using color thresholding and maximum entropy thresholding. The features are computed according to Bloom Richardson grading criteria. The decision tree classifier is used to classify input image into three group i.e., low grade, intermediate grade and high grade.

Much research has been accomplished in the area of drone landing and specifically pose estimation. While some of these works focus on sensor fusion using GPS, or GNSS, we propose a method that uses sensors, including four Time of Flight (ToF) range sensors and a monocular camera. However, when the descending platform is unstable, for example, on ships in the ocean, the uncertainty will grow, and the tracking will fail easily. We designed an algorithm that includes four ToF sensors for calibration and one for pose estimation. The landing process was divided into two main parts, the rendezvous and the final landing. Two important assumptions were made for these two phases. During the rendezvous, the landing platform movement can be ignored, while during the landing phase, the drone is assumed to be stable and waiting for the best time to land. The current research modifies the landing part as a stable drone and an unstable landing platform, which is a Stewart platform, with a mounted AprilTag. A novel algorithm for calibration was used based on color thresholding, a convex hull, and centroid extraction. Next, using the homogeneous coordinate equations of the sensors’ touching points, the focal length in the X and Y directions can be calculated. In addition, knowing the plane equation allows the Z coordinates of the landmark points to be projected. The homogeneous coordinate equation was then used to obtain the landmark’s X and Y Cartesian coordinates. Finally, 3D rigid body transformation is engaged to project the landing platform transformation in the camera frame. The test bench used Software-in-the-Loop (SIL) to confirm the practicality of the method. The results of this work are promising for unstable landing platform pose estimation and offer a significant improvement over the single-camera pose estimation AprilTag detection algorithms (ATDA).

Yield prediction is crucial for the management of harvest and scheduling wine production operations. Traditional yield prediction methods rely on manual sampling and are time-consuming, making it difficult to handle the intrinsic spatial variability of vineyards. There have been significant advances in automatic yield estimation in vineyards from on-ground imagery, but terrestrial platforms have some limitations since they can cause soil compaction and have problems on sloping and ploughed land. The analysis of photogrammetric point clouds generated with unmanned aerial vehicles (UAV) imagery has shown its potential in the characterization of woody crops, and the point color analysis has been used for the detection of flowers in almond trees. For these reasons, the main objective of this work was to develop an unsupervised and automated workflow for detection of grape clusters in red grapevine varieties using UAV photogrammetric point clouds and color indices. As leaf occlusion is recognized as a major challenge in fruit detection, the influence of partial leaf removal in the accuracy of the workflow was assessed. UAV flights were performed over two commercial vineyards with different grape varieties in 2019 and 2020, and the photogrammetric point clouds generated from these flights were analyzed using an automatic and unsupervised algorithm developed using free software. The proposed methodology achieved R2 values higher than 0.75 between the harvest weight and the projected area of the points classified as grapes in vines when partial two-sided removal treatment, and an R2 of 0.82 was achieved in one of the datasets for vines with untouched full canopy. The accuracy achieved in grape detection opens the door to yield prediction in red grape vineyards. This would allow the creation of yield estimation maps that will ease the implementation of precision viticulture practices. To the authors’ knowledge, this is the first time that UAV photogrammetric point clouds have been used for grape clusters detection.

Breast cancer is the common existing form of cancers amongst women. The automatic image analysis methods have an enormous potential to decrease the workload in a pathology laboratory. The grading of breast cancer histopathology images is used to find the level of breast cancer. The automatic grading of breast cancer histopathology images is a challenging task. In this paper a system for automatic detection of breast cancer grading of H&E stained histopathological images is presented. An image processing techniques such as preprocessing, segmentation, feature extraction and classification are used in this system. The segmentation of nuclei in H&E stained image is performed using color thresholding and maximum entropy thresholding. The features are computed according to Bloom Richardson grading criteria. The decision tree classifier is used to classify input image into three group i.e., low grade, intermediate grade and high grade. [ABSTRACT FROM AUTHOR]

Biopsy is a time consuming process and requires a great deal of skill and experience. Digital image processing techniques can play an important role in helping perform breast biopsies, especially of abnormal areas. In my research work, to understand the type of breast cancer cell and attempt to analyses the histopathological slides with my proposed method to identify cancer parts just using simple technique of isolation of insignificant portion of biopsy slide by color thresholding based segmentation. Many features used in the breast cancer detection and analysis of histopathology image are inspired by clinical pathologists as important for diagnosis and characterization. A large majority of these features are features of cell nuclei in biopsy image; as such, there is often the desire to segment the image into individual cell nuclei. In this paper, present an analysis of the utility of color thresholding method for segmentation of cancer cell nuclei for classification of H & E stained histopathology image of breast tissue. I am showing the cell level classification performance using these segmented nuclei in a benign versus malignant. Results indicate that very good segmentation and classification accuracies can be achieved with color thresholding based segmentation of cancer cell nuclei. The simplicity of algorithm is leads to less computational time. Thus, this approach is suitable for automated real time breast cancer detection and analysis tool. [ABSTRACT FROM AUTHOR]

Yield prediction is crucial for the management of harvest and scheduling wine production operations. Traditional yield prediction methods rely on manual sampling and are time-consuming, making it difficult to handle the intrinsic spatial variability of vineyards. There have been significant advances in automatic yield estimation in vineyards from on-ground imagery, but terrestrial platforms have some limitations since they can cause soil compaction and have problems on sloping and ploughed land. The analysis of photogrammetric point clouds generated with unmanned aerial vehicles (UAV) imagery has shown its potential in the characterization of woody crops, and the point color analysis has been used for the detection of flowers in almond trees. For these reasons, the main objective of this work was to develop an unsupervised and automated workflow for detection of grape clusters in red grapevine varieties using UAV photogrammetric point clouds and color indices. As leaf occlusion is recognized as a major challenge in fruit detection, the influence of partial leaf removal in the accuracy of the workflow was assessed. UAV flights were performed over two commercial vineyards with different grape varieties in 2019 and 2020, and the photogrammetric point clouds generated from these flights were analyzed using an automatic and unsupervised algorithm developed using free software. The proposed methodology achieved R2 values higher than 0.75 between the harvest weight and the projected area of the points classified as grapes in vines when partial two-sided removal treatment, and an R2 of 0.82 was achieved in one of the datasets for vines with untouched full canopy. The accuracy achieved in grape detection opens the door to yield prediction in red grape vineyards. This would allow the creation of yield estimation maps that will ease the implementation of precision viticulture practices. To the authors' knowledge, this is the first time that UAV photogrammetric point clouds have been used for grape clusters detection. [ABSTRACT FROM AUTHOR]