The virus SARS-CoV2, which causes coronavirus disease (COVID-19) has become a pandemic and spread to every inhabited continent. Given the increasing caseload, there is an urgent need augment clinical skills in order identify from among many mild cases few that will progress critical illness. We present first step towards building artificial intelligence (AI) framework, with predictive analytics (PA) capabilities applied real patient data, provide rapid decision-making support. COVID-19...

Enabled by the rapid development of virtual reality hardware and software, 360-degree video content has proliferated. From network perspective, transmission imposes significant challenges because it consumes 4 6χ bandwidth a regular with same resolution. To address these challenges, in this paper, we propose motion-prediction-based mechanism that matches to viewer needs. Ideally, if motion is perfectly known advance, could reduce consumption 80%. Practically, however, guarantee quality...

Colorectal cancer (CRC) is one of the most common types with a high mortality rate. Colonoscopy preferred procedure for CRC screening and has proven to be effective in reducing mortality. Thus, reliable computer-aided polyp detection classification system can significantly increase effectiveness colonoscopy. In this paper, we create an endoscopic dataset collected from various sources annotate ground truth location results help experienced gastroenterologists. The serve as benchmark platform...

360-degree video is on the cusp of going mainstream. Such videos have a large size (4-5⌉ regular one). Since each viewer has to wear Head-Mounted Display (HMD), screen sharing impossible when group watching same content. Multiple parallel video-streams will be required serve viewers along last mile (think family or friends Super Bowl in their HMDs). This would end up quickly choking entire network. In this paper, we present scheme optimize network bandwidth using motion-prediction-based...

To localize structural laryngeal lesions within digital flexible laryngoscopic images and to classify them as benign or suspicious for malignancy using state-of-the-art computer vision detection models.Cross-sectional diagnostic study SETTING: Tertiary care voice clinic METHODS: Digital stroboscopic videos, demographic clinical data were collected from patients evaluated a lesion. Laryngoscopic extracted videos manually labeled with bounding boxes encompassing the Four models employed...

This paper focuses on the problem of real-time detection and tracking a golf ball from video sequences. We propose an efficient effective solution by integrating object discrete Kalman model. For detection, three classical convolutional neural network based models are implemented, including Faster R-CNN, YOLOv3, YOLOv3 tiny. At stage, filter is employed to predict location previous observations. To increase accuracy speed, we use image patches rather than entire images for detection. In...

Flexible laryngoscopy is commonly performed by otolaryngologists to detect laryngeal diseases and recognize potentially malignant lesions. Recently, researchers have introduced machine learning techniques facilitate automated diagnosis using images achieved promising results. The diagnostic performance can be improved when patients' demographic information incorporated into models. However, the manual entry of patient data time-consuming for clinicians. In this study, we made first endeavor...

Abstract Aphid infestation poses a significant threat to crop production, rural communities, and global food security. While chemical pest control is crucial for maximizing yields, applying chemicals across entire fields both environmentally unsustainable costly. Hence, precise localization management of aphids are essential targeted pesticide application. The paper primarily focuses on using deep learning models detecting aphid clusters. We propose novel approach estimating infection levels...

Label assignment plays a significant role in modern object detection models. Detection models may yield totally different performances with label strategies. For anchor-based models, the IoU (Intersection over Union) threshold between anchors and their corresponding ground truth bounding boxes is key element since positive samples negative are divided by threshold. Early detectors simply utilize fixed for all training samples, while recent algorithms focus on adaptive thresholds based...

The Vision Transformer (ViT) leverages the Transformer's encoder to capture global information by dividing images into patches and achieves superior performance across various computer vision tasks. However, self-attention mechanism of ViT captures context from outset, overlooking inherent relationships between neighboring pixels in or videos. Transformers mainly focus on while ignoring fine-grained local details. Consequently, lacks inductive bias during image video dataset training. In...

Aphid infestations can cause extensive damage to wheat and sorghum fields spread plant viruses, resulting in significant yield losses agriculture. To address this issue, farmers often rely on chemical pesticides, which are inefficiently applied over large areas of fields. As a result, considerable amount pesticide is wasted without pests, while inadequate amounts with severe infestations. The paper focuses the urgent need for an intelligent autonomous system that locate spray within complex...

Aphid infestations are one of the primary causes extensive damage to wheat and sorghum fields most common vectors for plant viruses, resulting in significant agricultural yield losses. To address this problem, farmers often employ inefficient use harmful chemical pesticides that have negative health environmental impacts. As a result, large amount pesticide is wasted on areas without pest infestation. This brings attention urgent need an intelligent autonomous system can locate spray...

This paper focuses on the problem of online golf ball detection and tracking from image sequences. An efficient real-time approach is proposed by exploiting convolutional neural networks (CNN) based object a Kalman filter prediction. Five classical deep learning-based are implemented evaluated for detection, including YOLO v3 its tiny version, v4, Faster R-CNN, SSD, RefineDet. The performed small patches instead entire to increase performance detection. At stage, discrete employed predict...

Object Detection with Transformers (DETR) and related works reach or even surpass the highly-optimized Faster-RCNN baseline self-attention network architectures. Inspired by evidence that pure possesses a strong inductive bias leads to transformer losing expressive power respect depth, we propose architecture mitigatory mechanism applying possible direct mapping connections in mitigate rank collapse so as counteract feature expression loss enhance model performance. We apply this proposal...

Background Smoking and secondhand smoke (SHS) exposure rates are much higher in China than other countries. A smoke-free policy was implemented Xi'an, a city Shaanxi Province, China, on November 1, 2018. This study aimed to evaluate the effect of changes hospital admissions for acute ischemic heart disease (AIHD) stroke Xi'an. Methods All subjects had been hospitalized AIHD or from February 9, 2017 December 25, 2019 (study period: 150 weeks) six randomly selected public hospitals out 36...

Aphid infestations are one of the primary causes extensive damage to wheat and sorghum fields most common vectors for plant viruses, resulting in significant agricultural yield losses. To address this problem, farmers often employ inefficient use harmful chemical pesticides that have negative health environmental impacts. As a result, large amount pesticide is wasted on areas without pest infestation. This brings attention urgent need an intelligent autonomous system can locate spray...

3D point cloud classification requires distinct models from 2D image due to the divergent characteristics of respective input data. While clouds are unstructured and sparse, images structured dense. Bridging domain gap between these two data types is a non-trivial challenge enable model interchangeability. Recent research using Lattice Point Classifier (LPC) highlights feasibility cross-domain applicability. However, lattice projection operation in LPC generates with disconnected projected...

Vision Transformers have made remarkable progress in recent years, achieving state-of-the-art performance most vision tasks. A key component of this success is due to the introduction Multi-Head Self-Attention (MHSA) module, which enables each head learn different representations by applying attention mechanism independently. In paper, we empirically demonstrate that can be further enhanced overlapping heads MHSA. We introduce Multi-Overlapped-Head (MOHSA), where are overlapped with their...

Aphids are one of the main threats to crops, rural families, and global food security. Chemical pest control is a necessary component crop production for maximizing yields, however, it unnecessary apply chemical approaches entire fields in consideration environmental pollution cost. Thus, accurately localizing aphid estimating infestation level crucial precise local application pesticides. Aphid detection very challenging as each individual really small all aphids crowded together clusters....