Generative Deep Learning in Digital Pathology Workflows
یادگیری عمیق مولد در گردش کار آسیب شناسی دیجیتال-2021
Many modern histopathology laboratories are in the process of digitizing their workflows. Digitization of tissue images has made it feasible to research the augmentation or automation of clinical reporting and diagnosis. The application of modern computer vision techniques, based on deep learning, promises systems that can identify pathologies in slide images with a high degree of accuracy. Generative modeling is an approach to machine learning and deep learning that can be used to transform and generate data. It can be applied to a broad range of tasks within digital pathology, including the removal of color and intensity artifacts, the adaption of images in one domain into those of another, and the generation of synthetic digital tissue samples. This review provides an introduction to the topic, considers these applications, and discusses future directions for generative models within histopathology.
Automatic identification and quantification of dense microcracks in high-performance fiber-reinforced cementitious composites through deep learning-based computer vision
شناسایی و تعیین کمی ترکهای متراکم در کامپوزیت های سیمانی با عملکرد بالا با استفاده از دید رایانه ای مبتنی بر یادگیری عمیق-2021
High-performance fiber-reinforced cementitious composites (HPFRCCs) feature high mechanical strengths, crack resistance, and durability. Under excessive loading, HPFRCCs demonstrate dense microcracks that are difficult to identify using existing methods. This study presents a computer vision method for identification, quantification, and visualization of microcracks in HPFRCCs based on deep learning. The presented method integrates multiple deep learning models and computer vision techniques in a hierarchical architecture. The crack pattern (e.g., number, width, and spacing of cracks) are automatically determined from pictures without human intervention. This study shows that the presented method achieves an accuracy of 0.992 for crack detection and an accuracy finer than 50 μm (R2 > 0.984) for quantification of crack width when deep learning models are trained using only 200 pictures of HPFRCCs and 200 pictures of conventional concrete with incorporation of data augmentation. The presented method is expected to be also applicable to other materials featuring complex cracks.
Keywords: Computer vision | Crack detection | Crack quantification | Deep learning | High-performance fiber reinforced | cementitious composites (HPFRCC) | Microcrack
A fusing framework of shortcut convolutional neural networks
چارچوبی تلفیقی از شبکه های عصبی پیچشی میانبر-2021
Convolutional neural networks (CNNs) have proven to be very successful in learning task- specific computer vision features. To integrate features from different layers in standard CNNs, we present a fusing framework of shortcut convolutional neural networks (S-CNNs). This framework can fuse arbitrary scale features by adding weighted shortcut connections to the standard CNNs. Besides the framework, we propose a shortcut indicator (SI) of binary string to stand for a specific S-CNN shortcut style. Additionally, we design a learning algorithm for the proposed S-CNNs. Comprehensive experiments are conducted to compare its performances with standard CNNs on multiple benchmark datasets for different visual tasks. Empirical results show that if we choose an appropriate fusing style of shortcut connections with learnable weights, S-CNNs can perform better than standard CNNs regarding accuracy and stability in different activation functions and pooling schemes initializations, and occlusions. Moreover, S-CNNs are competitive with ResNets and can outperform GoogLeNet, DenseNets, Multi-scale CNN, and DeepID.© 2021 Elsevier Inc. All rights reserved.
Keywords: Convolutional neural networks | Computer vision | Shortcut connections
American Society of Biomechanics Early Career Achievement Award 2020: Toward Portable and Modular Biomechanics Labs: How Video and IMU Fusion Will Change Gait Analysis
انجمن آمریکایی بیومکانیک جایزه دستاورد اولیه شغلی 2020: به سوی آزمایشگاههای بیومکانیک قابل حمل و مدولار: چگونه تجزیه ویدئو و IMU تجزیه و تحلیل راه رفتن را تغییر می دهد-2021
The field of biomechanics is at a turning point, with marker-based motion capture set to be replaced by portable and inexpensive hardware, rapidly improving markerless tracking algorithms, and open datasets that will turn these new technologies into field-wide team projects. Despite progress, several challenges inhibit both inertial and vision-based motion tracking from reaching the high accuracies that many biomechanics applications require. Their complementary strengths, however, could be harnessed toward better solutions than those offered by either modality alone. The drift from inertial measurement units (IMUs) could be corrected by video data, while occlusions in videos could be corrected by inertial data. To expedite progress in this direction, we have collected the CMU Panoptic Dataset 2.0, which contains 86 subjects captured with 140 VGA cameras, 31 HD cameras, and 15 IMUs, performing on average 6.5 minutes of activities, including range of motion activities and tasks of daily living. To estimate ground-truth kinematics, we imposed simultaneous consistency with the video and IMU data. Threedimensional joint centers were first computed by geometrically triangulating proposals from a convolutional neural network applied to each video independently. A statistical meshed model parametrized in terms of body shape and pose was then fit through a top-down optimization approach that enforced consistency with both the video-based joint centers and IMU data. This sensor-dense dataset can be used to benchmark new methods that integrate a much sparser set of videos and IMUs to estimate both kinematics and kinetics in a markerless fashion.
Key words: markerless motion tracking | computer vision | inertial measurement units
Computer-vision classification of corn seed varieties using deep convolutional neural network
طبقه بندی بینایی ماشین انواع بذر ذرت با استفاده از شبکه عصبی پیچیده عمیق-2021
Automated classiﬁcation of seed varieties is of paramount importance for seed producers to maintain the purity of a variety and crop yield. Traditional approaches based on computer vision and simple feature extraction could not guarantee high accuracy classiﬁcation. This paper presents a new approach using a deep convolutional neural network (CNN) as a generic feature extractor. The extracted features were classiﬁed with artiﬁcial neural network (ANN), cubic support vector machine (SVM), quadratic SVM, weighted k-nearest-neighbor (kNN), boosted tree, bagged tree, and linear discriminant analysis (LDA). Models trained with CNN-extracted features demonstrated better classiﬁcation accuracy of corn seed varieties than models based on only simple features. The CNN-ANN classiﬁer showed the best performance, classifying 2250 test instances in 26.8 s with classiﬁcation accuracy 98.1%, precision 98.2%, recall 98.1%, and F1-score 98.1%. This study demonstrates that the CNN-ANN classiﬁer is an efﬁcient tool for the intelligent classiﬁcation of different corn seed varieties.© 2021 Elsevier Ltd. All rights reserved.
Keywords: Machine vision | Deep learning | Feature extraction | Non-handcrafted features | Texture descriptors
Evaluating Congou black tea quality using a lab-made computer vision system coupled with morphological features and chemometrics
ارزیابی کیفیت چای سیاه کنگو با استفاده از یک سیستم بینایی کامپیوتری ساخته شده در آزمایشگاه همراه با ویژگی های مورفولوژیکی و شیمی سنجی-2021
The feature of external shape in tea is a vital quality index that determines the rank quality of tea. The potential of a lab-made computer vision system (CVS) coupled with morphological features and chemometric tools is investigated for evaluating Congou black tea quality. First, Raw images of 700 tea samples from seven different quality grades are acquired using the CVS. The original images collected are processed by graying, binarization, and median de-noising. Then, six morphological parameters (viz. width, length, area, perimeter, length-width ratio, and rectangularity) from the samples are extracted by the shape segmentation of each tea leaf image, and the corresponding feature histogram is obtained. Finally, support vector machine (SVM) and least squares- support vector machine (LS-SVM) are utilized to build identification models based on the histogram distribution characteristic vectors. Three kernel methods (linear kernel, polynomial kernel, and radial basis function kernel) are compared for monitoring tea quality. The results show that the optimal LS-SVM model has a 12% higher correct discrimination rate (CDR) than the SVM model. The polynomial kernel LS-SVM model yields satisfactory classification results with the CDR of 100% based on selected six shape features in the calibration and prediction sets. This work demonstrates that it is feasible to discriminate Congou black tea quality using CVS technology along with morphological features and nonlinear chemometric methods. A new perspective on the sizes of morphological characteristics is proposed as an identifier of Congou black tea quality.
Keywords: Congou black tea | Computer vision system | Morphological features | Least squares-support vector machine | Kernel method
Predicting silicon, aluminum, and iron oxides contents in soil using computer vision and infrared
پیش بینی محتوای اکسیدهای سیلیکون ، آلومینیوم و آهن در خاک با استفاده از بینایی ماشین و مادون قرمز-2021
Silicon, aluminum, and iron oxides are very abundant in soil. Their quantification is important for soil classi- fication, which is a relevant information for the sustainable use and management of soils. In soil laboratories the determination of these oxides, using standard methods, is destructive, costly, laborious, and time consuming. This article presents two analytical methods to quantify SiO2, Al2O3, and Fe2O3 in soil samples using computer vision (COMPVIS) and mid-infrared spectroscopy (MIR). These two methods were developed using 52 soil samples from four states of Brazil. Digital images and MIR spectra were correlated with oxides contents quantified by atomic absorption spectroscopy (AAS) after acid digestion using three multivariate calibration methods: PLS, SPA-MLR, and LS-SVM. This the first time that soil image data has been correlated to silicon and aluminum oxides and the proposed method found excellent correlation values (r2 ranging from 0.95 to 0.99). With the exception of SiO2, MIR resulted in similar predictions to the COMPVIS method’s. LS-SVM presented r2 higher than 0.95 for all oxides estimates. The developed analyses are low cost, fast, and environmentally sustainables.1.
Keywords: SVM | MIA | oil chemistry | Green chemistry | Hematite | Sustainability
Compensating over- and underexposure in optical target pose determination
Compensating over- and underexposure in optical target pose determination-2021
Optical coded targets allow to determine the relative pose of a camera, on a metric scale, from one image only. Furthermore, they are easily and eﬃciently detected, opening to a wide range of applications in robotics and computer vision. In this work we describe the effect of pixel saturation and non-ideal lens Point Spread Function, causing the apparent position of the corners and the edges of the target to change as a function of the camera exposure time. This effect, which we call exposure bias, is frequent in over- or underexposed images and introduces a systematic error in the estimated camera pose. We propose an algorithm that is able to estimate and correct for the exposure bias exploiting speciﬁc geometric features of a common target design based on concentric circles. Through rigorous laboratory experiments carried out in a highly controlled environment, we demonstrate that the proposed algorithm is seven times more precise and three times more accurate in the target distance estimation than the algorithms available in the literature.© 2021 Elsevier Ltd. All rights reserved.
Keywords: Optical target | Target orientation | Image processing algorithm | Geometry | Ellipse fitting | Computer vision | Overexposure | Exposure compensation | Resection
Weight and volume estimation of poultry and products based on computer vision systems: a review
Weight and volume estimation of poultry and products based on computer vision systems: a review-2021
The appearance, size, and weight of poultry meat and eggs are essential for production economics and vital in the poultry sector. These external characteristics influence their market price and consumers’ preference and choice. With technological developments, there is an increase in the application and importance of vision systems in the agricultural sector. Computer vision has become a promising tool in the realtime automation of poultry weighing and processing systems. Owing to its noninvasive and nonintrusive nature and its capacity to present a wide range of information, computer vision systems can be applied in the size, mass, volume determination, and sorting and grading of poultry products. This review article gives a detailed summary of the current advances in measuring poultry products’ external characteristics based on computer vision systems. An overview of computer vision systems is discussed and summarized. A comprehensive presentation of the application of computer vision-based systems for assessing poultry meat and eggs was provided, that is, weight and volume estimation, sorting, and classification. Finally, the challenges and potential future trends in size, weight, and volume estimation of poultry products are reported.
Key words: classification | computer vision | egg | weight estimation | poultry product
Deep learning-based computer vision to recognize and classify suturing gestures in robot-assisted surgery
بینایی عمیق مبتنی بر یادگیری برای تشخیص و طبقه بندی حرکات بخیه در جراحی با کمک روبات-2021
Background: Our previous work classified a taxonomy of suturing gestures during a vesicourethral anastomosis of robotic radical prostatectomy in association with tissue tears and patient outcomes. Herein, we train deep learning-based computer vision to automate the identification and classification of suturing gestures for needle driving attempts.
Methods: Using two independent raters, we manually annotated live suturing video clips to label timepoints and gestures. Identification (2,395 videos) and classification (511 videos) datasets were compiled to train computer vision models to produce 2- and 5-class label predictions, respectively. Networks were trained on inputs of raw red/blue/green pixels as well as optical flow for each frame. Each model was trained on 80/20 train/test splits.
Results: In this study, all models were able to reliably predict either the presence of a gesture (identification, area under the curve: 0.88) as well as the type of gesture (classification, area under the curve: 0.87) at significantly above chance levels. For both gesture identification and classification datasets, we observed no effect of recurrent classification model choice (long short-term memory unit versus convolutional long short-term memory unit) on performance.
Conclusion: Our results demonstrate computer vision’s ability to recognize features that not only can identify the action of suturing but also distinguish between different classifications of suturing gestures. This demonstrates the potential to utilize deep learning computer vision toward future automation of surgical skill assessment.