Self-Adaptive Feature Transformation Networks for Object Detection in low luminance Images

Despite the recent improvement of object detection techniques, many of them fail to detect objects in low-luminance images. The blurry and dimmed nature of low-luminance images results in the extraction of vague features and failure to detect objects. In addition, many existing object detection methods are based on models trained on both sufficient- and low-luminance images, which also negatively affect the feature extraction process and detection results. In this article, we propose a framework called Self-adaptive Feature Transformation Network (SFT-Net) to effectively detect objects in low-luminance conditions. The proposed SFT-Net consists of the following three modules: (1) feature transformation module, (2) self-adaptive module, and (3) object detection module. The purpose of the feature transformation module is to enhance the extracted feature through unsupervisely learning a feature domain projection procedure. The self-adaptive module is utilized as a probabilistic module producing appropriate features either from the transformed or the original features to further boost the performance and generalization ability of the proposed framework. Finally, the object detection module is designed to accurately detect objects in both low- and sufficient- luminance images by using the appropriate features produced by the self-adaptive module. The experimental results demonstrate that the proposed SFT-Net framework significantly outperforms the state-of-the-art object detection techniques, achieving an average precision (AP) of up to 6.35 and 11.89 higher on the sufficient- and low- luminance domain, respectively.

An Efficient Tongue Segmentation Model Based on U-Net Framework

In the intelligently processing of the tongue image, one of the most important tasks is to accurately segment the tongue body from a whole tongue image, and the good quality of tongue body edge processing is of great significance for the relevant tongue feature extraction. To improve the performance of the segmentation model for tongue images, we propose an efficient tongue segmentation model based on U-Net. Three important studies are launched, including optimizing the model’s main network, innovating a new network to specially handle tongue edge cutting and proposing a weighted binary cross-entropy loss function. The purpose of optimizing the tongue image main segmentation network is to make the model recognize the foreground and background features for the tongue image as well as possible. A novel tongue edge segmentation network is used to focus on handling the tongue edge because the edge of the tongue contains a number of important information. Furthermore, the advantageous loss function proposed is to be adopted to enhance the pixel supervision corresponding to tongue images. Moreover, thanks to a lack of tongue image resources on Traditional Chinese Medicine (TCM), some special measures are adopted to augment training samples. Various comparing experiments on two datasets were conducted to verify the performance of the segmentation model. The experimental results indicate that the loss rate of our model converges faster than the others. It is proved that our model has better stability and robustness of segmentation for tongue image from poor environment. The experimental results also indicate that our model outperforms the state-of-the-art ones in aspects of the two most important tongue image segmentation indexes: IoU and Dice. Moreover, experimental results on augmentation samples demonstrate our model have better performances.

DSMNN-Net: A Deep Siamese Morphological Neural Network Model for Burned Area Mapping Using Multispectral Sentinel-2 and Hyperspectral PRISMA Images

Wildfires are one of the most destructive natural disasters that can affect our environment, with significant effects also on wildlife. Recently, climate change and human activities have resulted in higher frequencies of wildfires throughout the world. Timely and accurate detection of the burned areas can help to make decisions for their management. Remote sensing satellite imagery can have a key role in mapping burned areas due to its wide coverage, high-resolution data collection, and low capture times. However, although many studies have reported on burned area mapping based on remote sensing imagery in recent decades, accurate burned area mapping remains a major challenge due to the complexity of the background and the diversity of the burned areas. This paper presents a novel framework for burned area mapping based on Deep Siamese Morphological Neural Network (DSMNN-Net) and heterogeneous datasets. The DSMNN-Net framework is based on change detection through proposing a pre/post-fire method that is compatible with heterogeneous remote sensing datasets. The proposed network combines multiscale convolution layers and morphological layers (erosion and dilation) to generate deep features. To evaluate the performance of the method proposed here, two case study areas in Australian forests were selected. The framework used can better detect burned areas compared to other state-of-the-art burned area mapping procedures, with a performance of >98% for overall accuracy index, and a kappa coefficient of >0.9, using multispectral Sentinel-2 and hyperspectral PRISMA image datasets. The analyses of the two datasets illustrate that the DSMNN-Net is sufficiently valid and robust for burned area mapping, and especially for complex areas.

Terminal++ for Robot Researcher Using C#

Purpose: Sometimes our robot researcher needs a terminal program to exchange the data with the robot or automation device. Nevertheless, the readily available terminal program lacks some functionality that is most relevant to the researcher. We feel that a featured rich terminal program can handle lots of communication overhead for the researcher and relieve them from repetitive and time-consuming tasks. In mind for this, we researched and developed a utility program. We added extra features like automatic send, change dynamic data, etc., so our robot researcher can test the system communication better. In this paper, we demonstrated the utility program in detail. It is built using C#, which is under the Microsoft dot net framework. The code is uploaded to GitHub. Anyone can download and use it. It can be customized for their need. All used classes are available in .cs format. Design/Methodology/Approach: This is the software utility program built by the dot net framework of Microsoft visual studio. It has a graphical user interface (GUI) and some object classes. It has a serial and ethernet interface to test the channel. Once the medium is selected, the application will send whatever is written in the input text box. The Data sending may be an automatic or manual process. In manual mode, after typing the command, we need to press the “Enter” key to send the data. In automatic mode, it will send automatically within the preset interval. The transmit and receive content is displayed inside the list box. Findings/results: sometimes, our project goes into a critical phase. We need to have good tools to overcome the situation immediately. This is a helpful tool to trace the communication-related issue. Using this tool, we can observe the outgoing and incoming data traffic. The robot researcher can use it for their communication-related debug purposes. Originality/Value: Using this terminal program, our robot researcher will get lots of benefits where readily available utility programs cannot provide them. It has some unique features like automatic sending, changing dynamic content, etc. It has a serial and ethernet interface channel so that most of the device communication can be debugged through this interface software. It is entirely free and open source. Anyone can download and use it for personal as well as commercial purposes. Paper Type: Experiment-based Research.

Crystal Structure, Thermodynamic Properties and DFT Studies of 5,6-dimethyl-1H-benzo[d]imidazol-3-ium 3-((2,4-dioxo-1,5-dioxaspiro[5.5]undecan-3-ylidene)methyl) -2,4-dioxo-1,5-dioxaspiro[5.5]undecane Hydrate

A new 1,5-dioxaspiro[5.5] derivative coupled with a benzimidazole moiety: 5,6-dimethyl-1H-benzo[d]imidazol-3-ium 3-((2,4-dioxo-1,5-dioxaspiro[5.5]undecan-3-ylidene) methyl) -2,4-dioxo-1,5-dioxaspiro[5.5]undecane hydrate (DBH) was prepared. The crystal structure confirmed that it belongs to triclinic, P-1 space group. The title compound includes one (C19H21O8)− anion, one (C9H11N2)+ cation and one water molecule, which assembled into a 2D-net framework by O–H···O and N–H···O hydrogen bonds. The quantum chemical computations using the B3LYP/6-311G (d, p) basis level of theory reveal that the optimized geometric structure is suitable to study the molecule. The theoretically simulated FT-IR spectra and electronic spectra of DBH are compared with experimental data. The results show that the B3LYP/6-311g (d, p) method fits well with the molecular structure. In addition, the thermodynamic properties have also been studied to determine the nature of the DBH.

Improving the Analysis of DNA Sanger Sequencing Results: SeqBase Computer Program

Введение. Программное обеспечение, предоставляемое производителями автоматических генетических анализаторов, в большинстве случаев позволяет провести адекватный анализ результатов секвенирования ДНК по Сэнгеру для матриц с составом нуклеотидов, близким к эквивалентному. Однако для рассмотрения результатов секвенирования матриц, отличающихся неэквивалентным нуклеотидным составом, требуется проводить анализ электрофореграмм с сохранением информации об интенсивности сигналов флуоресценции. В особенности это касается секвенирования ДНК, модифицированной бисульфитом натрия. Цель: разработать и апробировать в практике научных исследований компьютерную программу для обеспечения адекватного анализа электрофореграмм секвенирования ДНК по Сэнгеру на основе бережного отношения к первичным данным и аккуратного определения базовых линий в спектральных каналах отдельных нуклеотидов. Методы. Программа SeqBase написана на языке C#, программная платформа .NET Framework 4.0, и выполняется в среде исполнения CLR (Common Language Runtime) для операционных систем семейства Windows. Адрес установочного пакета программы SeqBase: http://www.epigenetic.ru/projects/seqbase. Результаты. Разработана компьютерная программа, предназначенная для анализа первичных результатов секвенирования по Сэнгеру (хроматограмм капиллярного электрофореза), полученных на автоматических генетических анализаторах и представленных в файлах формата ABIF (*.ab1), обеспечивающая следующие возможности: 1) просмотр исходных электрофореграмм как в общем виде, так и раздельно по спектральным каналам; 2) кадрирование области анализа; 3) сглаживание сигналов; 4) ручная установка базовой линии по каждому из спектральных каналов; 5) сведение базовых линий по всем каналам; 6) ручная коррекция подвижности фрагментов ДНК в зависимости от типа флуоресцентной метки терминирующего нуклеотида. Апробация программы успешно проведена в рамках ряда исследований, результаты которых опубликованы в рецензируемых научных изданиях. Заключение. Использование программы SeqBase целесообразно для анализа результатов секвенирования по Сэнгеру матриц ДНК с неэквивалентным нуклеотидным составом, в особенности, модифицированных бисульфитом натрия, во избежание получения ложных результатов и для уточнения количественных оценок. Background. The software provided by the manufacturers of automatic genetic analyzers, in most cases, allows an adequate analysis of the results of Sanger DNA sequencing for templates with a nucleotide composition close to the equivalent. However, to consider the results of sequencing of templates with non-equivalent nucleotide composition, it is necessary to analyze electrophoregrams with preservation of primary information on the intensity of fluorescence signals. This is especially important for the sequencing of DNA modified with sodium bisulfite. Aim: to develop and validate in the practice of scientific research a computer program that ensures adequate analysis of electrophoregrams of Sanger DNA sequencing based on preservation of the primary data and on accurate determination of baselines in the spectral channels of individual nucleotides. Methods. The SeqBase program is written in C#, the programming platform .NET Framework 4.0, and runs in the CLR (Common Language Runtime) for Windows operating systems. SeqBase installation package address is http://www.epigenetic.ru/projects/seqbase. Results. A computer program has been developed designed to analyze the primary results of Sanger sequencing (chromatograms of capillary electrophoresis) obtained from automatic genetic analyzers and presented in files of the ABIF (*.ab1) format, which provides the following functions: 1) viewing the original electrophoregrams both in general form and separately by spectral channels; 2) cropping the area of analysis; 3) signal smoothing; 4) manual setting of the baseline for each of the spectral channels; 5) convergence of baselines on all channels; 6) manual correction of the mobility of DNA fragments depending on the type of fluorescent label of the terminating nucleotide. The program has been successfully tested in a number of studies, the results of which have been published in peer-reviewed scientific journals. Conclusion. The use of the SeqBase program is advisable for the analysis of the results of Sanger sequencing of DNA templates with non-equivalent nucleotide composition, especially those modified with sodium bisulfite, to avoid false results and to clarify quantitative estimates.

An Unsupervised Learning-Based Multi-Organ Registration Method for 3D Abdominal CT Images

Medical image registration is an essential technique to achieve spatial consistency geometric positions of different medical images obtained from single- or multi-sensor, such as computed tomography (CT), magnetic resonance (MR), and ultrasound (US) images. In this paper, an improved unsupervised learning-based framework is proposed for multi-organ registration on 3D abdominal CT images. First, the explored coarse-to-fine recursive cascaded network (RCN) modules are embedded into a basic U-net framework to achieve more accurate multi-organ registration results from 3D abdominal CT images. Then, a topology-preserving loss is added in the total loss function to avoid a distortion of the predicted transformation field. Four public databases are selected to validate the registration performances of the proposed method. The experimental results show that the proposed method is superior to some existing traditional and deep learning-based methods and is promising to meet the real-time and high-precision clinical registration requirements of 3D abdominal CT images.