This paper proposes a system for the forecasting and automated inspection of rice Bakanae disease (RBD) infection rates via drone imagery. The proposed synthesizes camera calibrations area calculations in optimal data domain to detect infected bunches classify culm numbers. Optimal heights angles identification were examined linear discriminant analysis gradient magnitude by targeting morphological features RBD Camera calibration calculation enabled distortion correction simultaneous image...

Software as a Service (SaaS) is the most prevalent service delivery mode for cloud systems. This paper surveys common security vulnerabilities and corresponding countermeasures SaaS. It primarily focused on work published in last five years. We observe current SaaS trends lack of sufficiently broad robust some area such Identity Access management due to growth applications.

A website appears in search results based on web indexing conducted by a engine bot (e.g., crawler). Some webpages do not want to be found easily because they include sensitive information. There are several methods prevent crawlers from database. However, such can still indexed malicious crawlers. Through this study, we explore paradox perspective new use of captchas for prevention. Captchas used converting words captchas. We have implemented the web-based captcha conversion tool our...

The number of malware has steadily increased as spread and evasion techniques have advanced. Machine learning contributed to making analysis more efficient by detecting various behavioral patterns. However, when analyzing large-scale datasets, through models both high temporal spatial complexity. In order address these problems, this work proposes a low-dimensional feature using histogram entropy prototype selection algorithm hyperrectangles. forms an L × 256 map according the preselected...

Platform as a service (PaaS) provides middleware resources to cloud customers. As demand for PaaS services increases, so do concerns about the security of PaaS. This paper discusses principal and integrity requirements, vulnerabilities corresponding countermeasures. We consider three core elements - multi-tenancy, isolation, virtualisation how they relate trends such user resource side-channel in multi-tenant environments, protection sensitive data recommend number effective procedures which...

IoT devices are being exploited as entry points for cyberattacks due to security weaknesses. malware variants have evolved a result of vulnerabilities in devices. This study investigates whether can be detected with var-ious types family. An opcode sequence represent its family characteristics by utilizing categories. Opcode categories divided into 6 or 11 depending on functions. Thus, identify intrinsic the which it belongs. By applying entropy histogram, 2D representation category visually...

Identifying amyloid-beta positive patients is crucial for determining eligibility Alzheimer's disease (AD) clinical trials and new disease-modifying treatments, but currently requires PET or CSF sampling. Previous MRI-based deep learning models predicting amyloid positivity, using only T1w sequences, have shown moderate performance. We trained to predict positivity evaluated whether multi-contrast inputs improve A total of 4,058 exams with MRI PET-based quantitative deposition were obtained...

The Android operating system is the world's largest smartphone platform. platform carries with it significant risk due to vast amounts of malware that target Newer methods involve use machine learning detect malicious software in a more general manner, as opposed conventional rely on prior knowledge pertaining particular family. This study explores Convolutional Neural Network and various tree-based processing feature engineered from binaries. Overall model accuracy was found be between 87%...

Recently, various studies have emerged that utilize hyperspectral imaging for crop growth analysis and early disease diagnosis. However, the challenge of using numerous spectral bands or finding optimal area remains a difficult problem. In this paper, we propose method searching optimized band based on deep learning hyper-spectral image. The proposed extracts RGB images within to segment background foreground through Vision Transformer-based Seformer. segmented results project onto each...

Motivation: AD patients must undergo repeated visits for amyloid and tau radiotracer imaging, leading to high costs dose concerns due PET's inability simultaneously acquire multiple radiotracers during a single session. Goal(s): Using PET/MRI scans, we used deep learning create separate PET images from simulated combined dual-tracer image. Approach: We amyloid-tau image by blending co-registered list-mode data employed 2.5D U-Net architecture effective separation. Results: Mixed-dose models,...

This article concerns the labeling guideline for health benefits of livestock products. In recent years, products with have emerged as a key market product industries. However, current regulation functional foods severely prohibits industries from attaching most claims to Also, manufacturers some difficulties in certain because lack guidelines on benefit Therefore, and scientists strongly demanded revision regulation, Appended Chart No. 14 provided by Article 52 (2) Enforcement Regulation...

This study analyzed consumers’ characteristics for health functional food intake. A consumer survey was conducted to look various factors influencing the consumption of food. Heckman selection model is used control sample bias. The probit results show that women, married person, smaller family members, and higher income were more likely purchase In addition, interest in disease history, number visits hospital lead age, income, longer intake period induce amount These could be utilized...

농업분야에 대한 인공지능 모델의 적용과 관련된 연구가 매우 활발히 이루어지고 있다. 본 논문에서 딥러닝 전경(foreground) 정보와 배경)background) 정보를 융합함으로써 적은 파라미터로도 딸기 질병을 효율적으로 분류하는 모델을 제안하였다. 제안하는 방법은 전경과 배경의 특징을 분리하고 두 부분의 특징사이의 관계를 포착하여 분류성능을 향상시키기 때문에 파라메터 수가 단순한 모델이더라도 효율적인 분류가 가능하다. 확보한 수의 딸기영상 데이터로 네트워크의 학습을 위해 데이터를 증강하였으며, ImageNet 학습된 가중치를 전이학습 후 미세조정을 수행하였다. 제안된 7개 클래스로 분류한 질병의 분류 문제에서 95.6%의 정확도를 나타내었다.

소나무 재선충은 한국과 일본, 중국을 포함한 동아시아 지역의 소나무산림에 막대한 피해를 주는 원인이며, 피해목의 조기 발견과 제거는 재선충 확산을 막는 효과적인 방법이다. 본 논문에서는 드론으로 촬영되고 처리된 다중분광 정사영상에서 딥러닝 분류에 의한 피해목 탐색방법을 제안한다. 제안된 방법은 학습영상 데이터가 많지 않다는 가정 하에서 ResNet18을 백본으로 하는 패치기반의 분류기를 구성하고, 5밴드(RGB, NIR, RedEdge)의 다중 분광 정사영상을 분류하고 그 결과를 heat map 형태로 만든다. 정사영상의 map은 후처리를 통해 근사적 위치와 분포를 확산해가는 모습을 관찰할 수 있게 하며, 추출해낼 수도 있다. 연구의 패치기반 분류기 성능은 97.06%의 정확도로 96.50%의 RGB 영상분석 보다는 다소 개선된 나타내었다.

This study suggests a generalized availability and safety evaluation model to evaluate the influences system"s fault tolerant capabilities depending on automatic detection function such as periodic testings. The conventional of deals only with self diagnostics, supposes that coverage diagnostics is always constant. But all methods could be degraded. For example, surveillance test has potential human errors or equipment errors, degradation built-in logics, testing facilities. suggested models...

Type Inference is used in programming languages to improve writability. In this paper, we will be looking more specifically at Local Variable (LVTI). For those unfamiliar with LVTI, also give an in-depth explanation of what it and how works. There a lot debate surrounding modern day languages. More specifically, whether the costs associated LVTI outweigh benefits. It has found its way into many higher-level including C#, C++, JavaScript, Swift, Kotlin, Rust, Go, etc. look usefulness...

This paper presents an approach for securing software application chains in cloud environments. We use the concept of workflow management systems to explain model. Our prototype is based on Kepler scientific system enhanced with security analytics package.

Transfer learning of an advanced deep framework, utilizing 3D U-Net pre-trained model, and trained on human brain MRI scans is proposed for extraction from other species. The network architecture successfully performed whole rhesus macaque MRIs automatically with high accuracy, reduced errors, lower computational cost. Options data augmentation different rates were also tested. Successful implementation automated would offer the potential to apply same strategy animal models.

Change blindness is a phenomenon where an individual fails to notice alterations in visual scene when change occurs during brief interruption or distraction. Understanding this specifically important for the technique that uses stimulus, such as Virtual Reality (VR) Augmented (AR). Previous research had primarily focused on 2D environments conducted limited controlled experiments 3D immersive environments. In paper, we design and conduct two formal user investigate effects of different...

The main factor affecting the decline in rice yield is damage caused by localized heavy rains or typhoons. method of analyzing lodging area difficult to obtain objective results based on visual inspection and judgment field surveys visiting affected area. it requires a lot time money. In this paper, we propose estimation diagnosis for areas using Vision Transformer-based Segformer RGB images, which are captured unmanned aerial vehicles. proposed estimates lodging, normal, background model,...

Many image database systems utilize content-based retrieval technology using perceptual hashing to facilitate comparison and retrieval. However, conventional functions generate different hash codes for the same domain even with minor changes, making accurate searches in impossible. To overcome these limitations, a method has been devised enable search images similar queries synthesizing visual feature information, including texture, shape, color of images. This paper proposes deephashing...