检索结果-内蒙古大学图书馆

Purities prediction in a manufacturing froth flotation plant: the deep learning techniques

NEURAL COMPUTING & APPLICATIONS 2020年第17期32卷 13639-13649页

作者： Pu, Yuanyuan Szmigiel, Alicja Apel, Derek B. Chongqing Univ Coll Resources & Safety Engn State Key Lab Coal Mine Disaster Dynam & Control Chongqing 400044 Peoples R China Univ Alberta Sch Min & Petr Engn Edmonton AB T6G 2R3 Canada

Accurate and timely investigation to concentrate grade and recovery is a premise of realizing automation control in a froth flotation process. This study seeks to use deep learning technologies modeling a manufacturing flotation process, forecasting the concentrate purities for iron and the waste silica. Considering the size and temporality of engineering data, we adopted a long short-term memory to form the core part of the deep learning model. To perform this process, 23 variables reflecting a flotation plant were monitored and collected hourly over a half year time span, then wrangled, split, and restructured for deep learning model use. A deep learning model encompassing a stacked long short-term memory architecture was designed, trained, and tested with prepared data. The model's performance on test data demonstrates the capability of our proposed model to predict real-time concentrate purities for iron and silica. Compared with a traditional machine model typified by a random forest model in this study, the proposed deep learning model is significantly more competent to model a manufacturing froth flotation process. Expected to lay a foundation for realizing automation control of the flotation process, this study should encourage deep learning in mineral processing engineering.

关键词： Froth flotation deep learning Long short-term memory Concentrate purity

来源：评论

学校读者我要写书评

暂无评论

DXM-TransFuse U-net: Dual cross-modal transformer fusion U-net for automated nerve identification

引用

COMPUTERIZED MEDICAL IMAGING AND GRAPHICS 2022年 99卷 102090-102090页

作者： Xie, Baijun Milam, Gary Ning, Bo Cha, Jaepyeong Park, Chung Hyuk George Washington Univ Sch Engn & Appl Sci Dept Biomed Engn 800 22nd St NWSci & Engn HallSuite 5000 Washington DC 20052 USA Childrens Natl Hosp 111 Michigan Ave NW Washington DC 20010 USA

Accurate nerve identification is critical during surgical procedures to prevent damage to nerve tissues. Nerve injury can cause long-term adverse effects for patients, as well as financial overburden. Birefringence imaging is a noninvasive technique derived from polarized images that have successfully identified nerves that can assist during intraoperative surgery. Furthermore, birefringence images can be processed under 20 ms with a GPGPU implementation, making it a viable image modality option for real-time processing. In this study, we first comprehensively investigate the usage of birefringence images combined with deep learning, which can automatically detect nerves with gains upwards of 14% over its color image-based (RGB) counterparts on the F2 score. Additionally, we develop a deep learning network framework using the U-Net architecture with a Transformer based fusion module at the bottleneck that leverages both birefringence and RGB modalities. The dual-modality framework achieves 76.12 on the F2 score, a gain of 19.6 % over single-modality networks using only RGB images. By leveraging and extracting the feature maps of each modality independently and using each modality's information for cross-modal interactions, we aim to provide a solution that would further increase the effectiveness of imaging systems for enabling noninvasive intraoperative nerve identification.

关键词： image segmentation Medical imaging Multi-modal fusion deep learning

来源：评论

学校读者我要写书评

暂无评论

Efficient Polyp Segmentation via Integrity learning

Efficient Polyp Segmentation via Integrity Learning

引用

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Ziqiang Chen Kang Wang Yun Liu Department of Biochemistry and Molecular Biology MOE Key Laboratory of Metabolism and Molecular Medicine School of Basic Medical Sciences Fudan University Shanghai China Guangdong MedicineAI Technology Co. Ltd. Guangdong China School of Basic Medical Sciences Fudan University Shanghai China

Accurate polyp delineation in colonoscopy is crucial for assisting in diagnosis, image-guided interventions, and treatments. However, current deep-learning approaches fall short due to integrity defects, which often manifest as inadequate segmentation of lesions. This paper introduces the integrity concept in polyp segmentation at both macro and micro levels, aiming to alleviate integrity deficiency. Specifically, the model should distinguish entire polyps at the macro level and identify all components within polyps at the micro level. Our Integrity Capturing Polyp Segmentation (IC-PolypSeg) network utilizes lightweight backbones and 3 key components for integrity ameliorating: 1) Pixel-wise feature redistribution (PFR) module captures global spatial correlations across channels in the final semantic-rich encoder features. 2) Cross-stage pixel-wise feature redistribution (CPFR) module dynamically fuses high-level semantics and low-level spatial features to capture contextual information. 3) Coarse-to-fine calibration module combines PFR and CPFR modules to achieve precise boundary detection. Extensive experiments on 5 public datasets demonstrate that the proposed IC-PolypSeg outperforms 8 state-of-the-art methods in terms of higher precision and significantly improved computational efficiency with lower computational consumption. IC-PolypSeg-EF0 employs 300 times fewer parameters than PraNet while achieving a real-time processing speed of 235 FPS. Importantly, IC-PolypSeg reduces the false negative ratio on five datasets, meeting clinical requirements.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Smart Infant Monitoring System Using Computer Vision and AI

Smart Infant Monitoring System Using Computer Vision and AI

引用

IEEE International Conference on Electro-Information Technology

作者： Gurpreet Singh Abhishek Raj Shekhar Xinrui Yu Jafar Saniie Department of Electrical and Computer Engineering Embedded Computing and Signal Processing Research Laboratory Illinois Institute of Technology Chicago IL U.S.A.

The new era of technology is being greatly influenced by the field of artificial intelligence. Computer vision and deep learning have become increasingly important due to their ability to process vast amounts of data and provide insights and solutions in a variety of fields. Computer vision, deep learning and signal analysis have been used in a growing number of applications and services including smart devices, image, and speech recognition, healthcare, etc., one such device is an infant monitoring system. It monitors the daily activities of the infant such as their sleeping patterns, sounds, and movements. In this paper, deep learning and computer vision libraries were used to develop algorithms to detect whether the infant was in any uncomfortable situation such as sleeping on its back, face being covered and whether the infant was awake. The smart infant monitoring system detects the infant's unsafe resting situation in real time and sent immediate alerts to the caretaker's device. This paper presents the design flow of a smart infant monitoring system consisting of a night vision camera, a Jetson Nano, and a Wi-Fi internet connection. The pose estimation and awake detection algorithms were developed and tested successfully for different infant resting/sleeping situations. The smart infant monitoring system provides significant benefits for safety and an improved understanding of infants' sleep patterns and behavior.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Utilizing image processing and the YOLOv3 Network for real-time Traffic Light Control

JOURNAL OF ENGINEERING

引用

JOURNAL OF ENGINEERING 2023年第1期2023卷

作者： Altamirano, S. Francisco Segura Cardenas, Diana Castro M. Montes, Ayax Sifuentes M. Cabrera, Lucia Chaman I. Puelles, Esther Lizana Y. Coronel, Angel Rojas M. Rodriguez, Oscar de la Cruz M. Romero, Luis Lara A. Univ Nacl Pedro Ruiz Gallo Lambayeque Peru Univ Nacl Piura Piura Peru Univ Senor Sipan Chiclayo Peru Antenor Orrego Private Univ Trujillo Peru Natl Univ Trujillo Trujillo Peru

In this study, different strategies used to count vehicles and people in different image areas at a street intersection were analyzed to obtain counts at appropriate times suitable for real-time control of a traffic light. To achieve this, video recordings of cameras placed at the intersection were used to test and verify image processing algorithms and deep learning using the YOLOv3 network implemented on a 4 GB RAM Jetson Nano card. We counted the vehicles and people that stopped and crossed the polygons to delimit the different areas of interest, with a maximum error of +/- 2 in the validation tests for all cases. In addition, as a strategy, we combined the images from both cameras into a single one, thereby allowing us to make a single detection and subsequently determine if they are inside or outside the polygons used in separating the areas of interest with the respective counts. Furthermore, this enabled us to obtain information on vehicles and people stopped and crossing in a time of 0.73 s on average. Hence, it was established that the inclusion of the control algorithm is appropriate for real-time control of traffic lights.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Low-Latency Label-Free image-Activated Cell Sorting Using Fast deep learning and Ai Inferencing

SSRN

引用

SSRN 2022年

作者： Tang, Rui Xia, Lin Gutierrez, Bien Gagne, Ivan Munoz, Adonary Eribez, Korina Jagnandan, Nicole Chen, Xinyu Zhang, Zunming Waller, Lauren Alaynick, William Cho, Sung Hwan An, Cheolhong Lo, Yu-Hwa Department of Electrical and Computer Engineering University of California La Jolla San DiegoCA92093 United States NanoCellect Biomedical Inc. San DiegoCA92121 United States Department of Bioengineering University of California La Jolla San DiegoCA92093 United States

Classification and sorting of cells using image-activated cell sorting (IACS) systems can bring significant insight to biomedical sciences. Incorporating deep learning algorithms into IACS enables cell classification and isolation based on complex and human-vision uninterpretable morphological features within a heterogeneous cell population. However, the limited capabilities and complicated implementation of deep learning–assisted IACS systems reported to date hinder the adoption of the systems for a wide range of biomedical research. Here, we present image-activated cell sorting by applying fast deep learning algorithms to conduct cell sorting without labeling. The overall sorting latency, including signal processing and AI inferencing, is less than 3 milliseconds, and the training time for the deep learning model is less than 30 minutes with a training dataset of 20,000 images. Both values set the record for IACS with sorting by AI inference. The system provides a compact, sterile, low-cost, label-free, and low-latency cell sorting solution based on real-time AI inferencing and fast training of the deep learning model. © 2022, The Authors. All rights reserved.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Detection of typical Pathologies and Differential Treatments in Dental Panoramic X-Rays based on deep Convolutional Neural Network

Detection of typical Pathologies and Differential Treatments...

引用

International Conference on Control Communication & Computing India (ICCC)

作者： Athira Thulaseedharan Lal Priya P S Department of Electrical Engineering College of Engineering Trivandrum Thiruvananthapuram Kerala India

Artificial Intelligence (AI) is gaining traction in medical imaging. There are a plethora of potential applications, covering every stage of the medical imaging life cycle from image creation to diagnosis to outcome prognosis. The use of Artificial Intelligence (AI) and Machine learning is having a rising influence in the field of dentistry and is improving the growth of digital tools and technology, with a wide range of applications in cosmetic dental treatments and treatment planning. deep learning has been able to match and improve human performance in fields such as image processing for extremely complex tasks such as object detection models with classification and prediction capabilities, as well as the identification of numerous dental diseases and their differential treatments. The lack of sufficiently large, curated, and representative training data with expert labeling (eg, annotations) is one of the main barriers to the development and clinical application of AI algorithms. This pilot study used a dataset of 664 dental panoramic X-ray images to build a model that can detect dental diseases and the differential treatments present on a full dental X-ray. After being manually labeled for 9 different classes, these images were used for training. The deep neural network used in this study is You Only Look Once (YOLO) version 6, a one-step search method primarily used for real-time object detection. Three classes of dental diseases and six classes of differential treatments were identified with the aid of bounding boxes. Results show that this strategy was able to attain a mean average precision (mAP) of 70.76% by training with this smaller number of images, which shows tremendous potential for dentists and radiologists.

关键词：

来源：评论

学校读者我要写书评

暂无评论

CAN3D: Fast 3D medical image segmentation via compact context aggregation

引用

MEDICAL image ANALYSIS 2022年 82卷 102562页

作者： Dai, Wei Woo, Boyeong Liu, Siyu Marques, Matthew Engstrom, Craig Greer, Peter B. Crozier, Stuart Dowling, Jason A. Chandra, Shekhar S. Univ Queensland Sch Informat Technol & Elect Engn Brisbane Qld Australia CSIRO Australian E Hlth Res Ctr Canberra ACT Australia Univ Newcastle Callaghan NSW Australia

Direct automatic segmentation of objects in 3D medical imaging, such as magnetic resonance (MR) imaging, is challenging as it often involves accurately identifying multiple individual structures with complex geometries within a large volume under investigation. Most deep learning approaches address these challenges by enhancing their learning capability through a substantial increase in trainable parameters within their models. An increased model complexity will incur high computational costs and large memory requirements unsuitable for real-time implementation on standard clinical workstations, as clinical imaging systems typically have low-end computer hardware with limited memory and CPU resources only. This paper presents a compact convolutional neural network (CAN3D) designed specifically for clinical workstations and allows the segmentation of large 3D Magnetic Resonance (MR) images in real-time. The proposed CAN3D has a shallow memory footprint to reduce the number of model parameters and computer memory required for state-of-the-art performance and maintain data integrity by directly processing large full-size 3D image input volumes with no patches required. The proposed architecture significantly reduces computational costs, especially for inference using the CPU. We also develop a novel loss function with extra shape constraints to improve segmentation accuracy for imbalanced classes in 3D MR images. Compared to state-of-the-art approaches (U-Net3D, improved U-Net3D and V-Net), CAN3D reduced the number of parameters up to two orders of magnitude and achieved much faster inference, up to 5 times when predicting with a standard commercial CPU (instead of GPU). For the open-access OAI-ZIB knee MR dataset, in comparison with manual segmentation, CAN3D achieved Dice coefficient values of (mean = 0.87 +/- 0.02 and 0.85 +/- 0.04) with mean surface distance errors (mean = 0.36 +/- 0.32 mm and 0.29 +/- 0.10 mm) for imbalanced classes such as (femoral and tibial) ca

关键词： Semantic segmentation Magnetic resonance imaging deep learning Prostate and knee 3D

来源：评论

学校读者我要写书评

暂无评论

Liquid Petroleum Hydrocarbon Ocean Coastal Water Pollution Identification Using deep Neural Network

Liquid Petroleum Hydrocarbon Ocean Coastal Water Pollution I...

引用

International Conference on Advanced Computing and Communication Systems (ICACCS)

作者： A. Vinesh Maria Anto B. Venkat Eswar Thilagavathi C N. Subash K. Riza Thoufiq Department of Information Technology M. Kumarasamy College of Engineering Karur India

All marine life is put at peril when oil spills damage the marine ecosystem. Sea otters, and birds lose their natural protection against the cold and wet because oil breaks down their natural insulation and leads to hypothermia, therefore oil must be cleaned. Some of the effective ways of cleaning oils can include using chemical solvents, skimmers and dispersants and other methods. But all these methods have its own disadvantages when oils get dispersed from one place to another place in the ocean. Many methods like random-walk method, noise removal, and dark spot detection are used. Compared to these methods Convolutional Neural Network (CNN) has high computing and processing ability, the CNN architecture that may be used to semantically split SAR pictures into several categories. The suggested CNN is optimized for use as there is no physical contact required and also not implementing the sensors in ocean that could be costly. The advantages of the present Convolutional Neural Network (CNN) are CNNs have been shown to achieve high accuracy in a variety of computer vision tasks, including image classification, object detection, and segmentation. It has the Ability to learn complex features, Robustness to noise and real-time performance while other methods fail to do. The oil is detected the alert message is sent as an SMS to the administrator.

关键词： deep learning image segmentation Solvents Oils Oceans Sea measurements real-time systems

来源：评论

学校读者我要写书评

暂无评论

Optimal deep learning based fusion model for biomedical image classification

引用

EXPERT SYSTEMS 2022年第3期39卷

作者： Mansour, Romany F. Alfar, Nada M. Abdel-Khalek, Sayed Abdelhaq, Maha Saeed, Rashid A. Alsaqour, Raed New Valley Univ Fac Sci Dept Math El Kharga Egypt Taif Univ Coll Sci Dept Math At Taif Saudi Arabia Sohag Univ Fac Sci Dept Math Sohag Egypt Princess Nourah Bint Abdulrahman Univ Coll Comp & Informat Sci Dept Informat Technol Riyadh Saudi Arabia Saudi Elect Univ Coll Comp & Informat Dept Informat Technol Riyadh Saudi Arabia

Automated examination of biomedical signals plays a vital role to diagnose diseases and offers useful data to several applications in the areas of physiology, sports medicine, and human-computer interface. The latest advancements in Artificial Intelligence (AI) have the ability to manage and analyse enormous biomedical datasets resulting in clinical decision making and real time applications. At the same time, Colorectal cancer (CRC) is the third most deadly disease affecting people over the globe. The utilization of AI techniques for the earlier identification of CRC has gained significant interest among the research communities. Therefore, this paper presents a novel AI based fusion model for CRC disease diagnosis and classification, named AIFM-CRC. The presented AIFM-CRC model primarily undergoes Gaussian filtering based noise removal and contrast enhancement as a preprocessing stage. In addition, a fusion based feature extraction process takes place where the SIFT based handcrafted features and Inception v4 based deep features are fused together. Besides, whale optimization algorithm tuned deep support vector machine model is employed as a classification technique to determine the existence of CRC. In order to highlight the proficient results analysis of the AIFM-CRC model, a comprehensive simulation analysis takes place. The resultant experimental values pointed out the betterment of the AIFM-CRC model by accomplishing a maximum accuracy of 96.18%.

关键词： artificial intelligence biomedical signals colorectal cancer deep learning fusion model image processing

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：