Search

We aimed to automatically detect and segment mediastinal lymph nodes, and to establish an objective method and reliable new response evaluation criteria to monitor the effectiveness of cancer treatment. The image processing techniques were applied and developed 3D measurement of the mediastinal lymph nodes based on automatic settings when segmenting the lymph node image. A repeatable and consistent lymph node evaluation system was created based on such features as the position of occurrence, grayness, and number of serial sections of lymph nodes. A total of 200 lymph node samples from Tri-Service General Hospital, Taiwan, were examined for statistical analysis. The proposed approach used weighted k-nearest neighbors for classification, achieving superior results with an accuracy and specificity of 97.5% and 99.4%, respectively. The volume of the lymph nodes was used as the reference index for tumor invasiveness evaluation. The error in the lymph node volume was 1.71% according to the verification results. Receiver operating characteristic (ROC) curves for each analysis were constructed and the area under the curve (AUC) was calculated with histopathology diagnosis as outcome for determining the optimal volume threshold of benign and malignant lymph nodes. It was observed that the lymph node volume was highly correlated with tumor invasion (p-value was less than 0.05). The experiment showed that the volume for the area under the ROC curve was 0.90 of tumor invasion evaluation. The lymph node volume was most effective in predicting tumor invasiveness, with the value 798.53 mm3 used as the standard for judging benignity and malignancy. [ABSTRACT FROM AUTHOR]

• Plant leaf disease recognition. • Detection of Tomato leaf disease detection. • Ensemble classification for plant leaf disease recognition. • Diversity in ensemble classification. Plant health and accurate and timely diagnosis of plant diseases are of great importance for human health and the economy of the farmers in the world. Therefore, researchers in the community are looking for fast and accurate methods for diagnosing and classifying different types of plants and their diseases. However, analyzing and classifying plant diseases can be challenged due to various diseases and farm conditions. This study aims to analyze and classify 13 classes of tomato disease in the farm and laboratory conditions with high accuracy and speed through ensemble classification. Therefore, 260 ensemble classifiers were designed using various preprocessing, different feature extraction, and different classifiers. Then, we compared the effectiveness of those ensemble methods to reach the optimal ensemble classifiers. Two databases, including the Plantvillage and Taiwan tomato leaves, were used for evaluating the accuracy and precision of the proposed method in the laboratory and field conditions with different challenges. The best ensemble classifier was able to classify based on the conditions of the shadow, brightness changes, disease similarity, background clutter, multiple leaves, and different textures with 95.98% accuracy. Moreover, a comparison was made between the several Deep learning models and 260 proposed ensemble models in the proposed method. In addition, the performance of the proposed method was better compared to the state-of-the-art Deep learning model. [ABSTRACT FROM AUTHOR]

Abstract: We use the two-dimensional (2D) electrical resistivity imaging (ERI) method to detect the contact between sediment and bedrock in a modern fluvial system. We performed eight 2D electrical resistivity surveys along the Peikang River in Kuohsing area, central Taiwan, with a pole–pole array of 1-m electrode spacing to produce a spatial resolution of about 1 m. The resistivity data were inverted into subsurface electrical structures using the least-squares inversion techniques. Then, the contact between the bedrock and sediment was found by using an image processing technique called Laplacian edge detection (LED), which represents an objective approach in interpreting various geophysical images. Numerical modeling demonstrates that isoresistivity lines are not in good agreement with the hypothetical interfaces between the bedrock and sediment, but the zero-lines of the Laplacian operation are strikingly accurate. Our results indicate a well-defined boundary in the resistivity structure that can be used to estimate the quantity of sediments covering bedrock, thus highlighting the utility of this technique in studies of landscape evolution, sediment transport, and sediment budgets. [Copyright &y& Elsevier]

The current practice of traditional in-class lecture for learning computer science (CS) in the high schools of Taiwan is in need of revamping. Teachers instruct on the use of commercial software instead of teaching CS concepts to students. The lack of more suitable teaching materials and limited classroom time are the main reasons for the software-based teaching practices. This study proposes a pre-class video viewing of the lecture content plus in-class supervised laboratory work as a model that can overcome the above problems. Experiments on teaching image processing concepts at two local high schools have been conducted. The experimental results show that the proposed model is a feasible and effective way for the teaching and learning of image processing. The new approach also facilitates student–student and student–teacher interactions in the classroom. Finally, an improvement in the understanding of the subject by the students and an increased interest toward learning CS are also evident. [Copyright &y& Elsevier]

Abstract: This work investigates atmospheric visibility by employing two statistical processes for high-frequency extraction: Sobel operator and fast Fourier transform (FFT). In replacing the traditional measurement methods for atmospheric visibility, the new approaches can provide accurate digital data urban visibility by establishing the numerical indices. The procedure is illustrated as follows. Digital images of urban areas in Kaohsiung, a city at the south of Taiwan, are analyzed according to brightness. High-frequency components of the image are extracted to calculate the index values by employing the Sobel operator and FFT. Finally, the correlation between each index and the visual range estimated by trained investigators are evaluated. A good correlation between two indices and the values obtained by visual investigation is affirmed by correlation coefficients (), 0.8139 and 0.7797, respectively. Furthermore, both indices are highly correlated with each other (). Convenient transmission and the exchangeability of digital images of the real-time landscape facilitate the publication of results on the world wide web (WWW). [Copyright &y& Elsevier]

In the paper machine vision, laser interferometer and coordinate measuring machine (CMM) are combined to develop a vision inspection system. The measurement capability of the developed system is investigated by measuring the distances between the lines on a standard line scale. The vision camera is used to replace the probe of the CMM to take the images of the interested lines on a line scale at two different positions. Meanwhile, the displacement of the CCD camera is measured using laser interferometer. Using subpixel edge localization and outlier-excluding least-squares regression, the distance between two interested lines is computed under an image plane coordinate system. By adding the displacement of the CCD camera measured using laser interferometer, the line space can be determined. Experiments have been performed repeatedly to measure the line space on the 1.00 and 300.0 mm line scales. Results indicate that the measured data only have a little fluctuation and are close to those obtained by the NML (National Measurement Laboratory, Taiwan). [Copyright &y& Elsevier]

Detecting road damage quickly and accurately facilitates the ability of road-maintenance agencies to make timely repairs to road surfaces, maintain optimal road conditions, optimize transportation safety, and minimize transportation costs. An extensive evaluation of eight deep-learning-based road-damage detection models was conducted in this study. Each model was trained on 9493 images sourced from multiple databases. The 16165 instances of road damage in these images were categorized into five types of damage, including longitudinal crack, horizontal crack, alligator damage, pothole-related crack, and line blurring. Two experiments were conducted that identified two models, single shot multi-box detector (SSD) Inception V2 and faster region-based convolutional neural networks (R-CNN) Inception V2, as providing the best balance of road-damage-detection accuracy and image processing time. These experiments demonstrated that increasing the diversity of image sources improved road-damage-detection model performance. In addition to combining data images from different sources with consistently relabeled damage instances, this study released road-damage image data from the road maintenance agency in Zhubei, Hsinchu County, Taiwan for research and other uses, increasing the limited amount of published image data sources and positively impacting future scholarly research into road damage detection. [ABSTRACT FROM AUTHOR]