Article
Energy & Fuels
Zengyi Lyu, Xiaowei Jia, Yao Yang, Keqi Hu, Feifei Zhang, Gaofeng Wang
Summary: The paper proposes a deep learning model combining CNN and LSTM to detect combustion instability using high-speed flame image sequences. The model achieves superior performance under various combustion conditions in swirl chamber, with high accuracy and short processing time per frame.
Article
Mathematics
Yihang Zhang, Yunsick Sung
Summary: With the development of artificial intelligence, techniques such as machine learning, object detection, and trajectory tracking have been applied to detect traffic accidents using closed-circuit television (CCTV). This paper proposes a method that generates and considers object trajectories using influence maps and a convolutional neural network (CNN) to improve the detection of traffic accidents. The proposed method achieved remarkable performance improvement compared to methods that only rely on CNN-based detection.
Article
Computer Science, Artificial Intelligence
Pravendra Singh, Pratik Mazumder, Mohammed Asad Karim, Vinay P. Namboodiri
Summary: This study introduces an efficient method for calibrating the outputs of convolutional layers, named Accuracy Booster, which achieves higher performance by channel-wise calibration with minimal extra parameters and computation. Experimental results demonstrate that this method outperforms existing techniques on multiple datasets and architectures, and also generalizes well to various tasks.
Article
Biochemistry & Molecular Biology
Rimsha Asad, Saif Ur Rehman, Azhar Imran, Jianqiang Li, Abdullah Almuhaimeed, Abdulkareem Alzahrani
Summary: Brain tumors can cause damage to the brain and surrounding tissues, blood vessels, and nerves if not treated promptly. Early detection is crucial to prevent fatal outcomes. Manual detection is challenging due to variations in tumor characteristics, thus an automatic system using a deep convolutional neural network is proposed. The system achieved high accuracy on the brain-tumor dataset and outperformed baseline methods, demonstrating its potential for application in other diseases.
Article
Computer Science, Artificial Intelligence
Miguel Cardenas-Montes
Summary: Outlier detection in deep learning involves automatically extracting features from raw data to identify anomalies, eliminating the need for extensive feature engineering. Urban areas deploy sensor networks to monitor air quality variables, which can be treated as time series or maps for machine learning algorithms to detect outliers.
Article
Computer Science, Information Systems
Jae-Min Lee, Min-Seok Seo, Dae-Han Kim, Sang-Woo Lee, Jong-Chan Park, Dong-Geol Choi
Summary: In this work, the authors propose a Split-and-Share Module (SSM), which splits given features into parts and shares them among multiple sub-classifiers, in order to improve the performance of image classification tasks and identify structural characteristics within the features. SSM can be easily integrated into various architectures and has been validated to show significant improvements over baseline architectures.
Article
Computer Science, Information Systems
Jose P. Amorim, Pedro H. Abreu, Joao Santos, Marc Cortes, Victor Vila
Summary: This study introduces a new approach for evaluating saliency map methods, which have various applications in medical image classification. By introducing natural perturbations and studying their impact on evaluation metrics adapted from saliency models, the effectiveness of this method is validated. The results show that this approach accurately evaluates the reliability of saliency map methods and has potential for application in other medical imaging tasks.
INFORMATION PROCESSING & MANAGEMENT
(2023)
Article
Automation & Control Systems
Shumei Chen, Jianbo Yu, Shijin Wang
Summary: This study proposes a one-dimensional convolutional neural network model optimized by reinforcement learning-based neural architecture search for multivariate process control, which shows excellent performance in detecting and recognizing process faults.
Article
Thermodynamics
Taekeun Yoon, Seon Woong Kim, Hosung Byun, Younsik Kim, Campbell D. Carter, Hyungrok Do
Summary: This article presents a deep learning strategy for fast and accurate gas property measurements using flame emission spectroscopy (FES). Denoising convolutional neural networks (CNN) are used to enhance the signal-to-noise ratio (SNR) of the short-gated spectrum and improve the accuracy of property estimation.
COMBUSTION AND FLAME
(2023)
Article
Computer Science, Artificial Intelligence
Caijuan Shi, Weiming Zhang, Changyu Duan, Houru Chen
Summary: This paper proposes a pooling-based feature pyramid network to enhance salient object detection performance. By designing U-shaped feature pyramid modules, pyramid pooling refinement modules, and channel attention modules, rich semantic information from high-level and low-level features can be captured effectively to improve detection accuracy.
IMAGE AND VISION COMPUTING
(2021)
Review
Computer Science, Artificial Intelligence
Luis Guarda, Juan E. Tapia, Enrique Lopez Droguett, Marcelo Ramos
Summary: This paper presents a Deep Learning-based method for drowsiness detection using CapsNet and spectrogram images of EEG signals. The proposed CapsNet model outperforms the traditional Convolutional Neural Network in terms of accuracy and sensitivity. This method is effective for handling small amounts of data and biomedical signals.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Zhonghong Ou, Zhaofengnian Wang, Fenrui Xiao, Baiqiao Xiong, Hongxing Zhang, Meina Song, Yan Zheng, Pan Hui
Summary: With the popularity of 5G networks and IoT applications, real-time environmental awareness becomes crucial. However, small object detection still faces challenges due to limited scales and low detection accuracy. To address these issues, the proposed AD-RCNN employs dynamic region proposal network, visual attention scheme, and adaptive dynamic training module. Experimental results demonstrate that AD-RCNN outperforms existing methods in terms of mAP and FPS.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Shi Luo, Xiongfei Li, Xiaoli Zhang
Summary: The goal of this research is to improve face detection performance by extending the sampling range of face aspect ratio. We propose a Wide Aspect Ratio Matching (WARM) strategy and a feature enhancement module called Receptive Field Diversity (RFD) to capture more representative features of extreme aspect ratio faces. Extensive experiments show the effectiveness of our method on various benchmarks.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Ashish Semwal, Narendra D. Londhe
Summary: Pain assessment is crucial for medical diagnosis and treatment. Researchers have been focusing on objective methods, such as a fully automated pain assessment system based on facial expressions. Joint learning from both appearance and shape-based features results in a more robust pain assessment model.
APPLIED SOFT COMPUTING
(2021)
Article
Computer Science, Information Systems
Qiuchen Zhu, Tran Hiep Dinh, Manh Duong Phung, Quang Phuc Ha
Summary: Drone imagery is increasingly used in automated inspection for infrastructure surface defects, with a proposed deep learning approach utilizing hierarchical convolutional neural networks and a contrast-based autotuned thresholding algorithm. Experimental results show that the proposed technique outperforms existing methods on various tested datasets, especially in crack detection, demonstrating its effectiveness in surface defect inspection.
Article
Computer Science, Information Systems
Zhen Zhong, Wanlin Gao, Abdul Mateen Khattak, Minjuan Wang
MULTIMEDIA TOOLS AND APPLICATIONS
(2020)
Article
Computer Science, Theory & Methods
Zhen Zhong, Minjuan Wang, Wanlin Gao, Lihua Zheng
Summary: A novel infrared and visible image fusion method, NSCT-GF, was proposed for pig-body segmentation and temperature detection, achieving higher segmentation accuracy and efficiency compared to conventional methods. Experimental results demonstrated significant improvements in segmentation accuracy and efficiency, laying a foundation for accurately measuring the temperature of pig-body.
MULTIDIMENSIONAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Computer Science, Information Systems
Zhen Zhong
Summary: The paper introduces a new multisource fusion algorithm, MCNNFuse, for extracting pig-body shape and temperature features. The algorithm fuses visible and infrared images using convolutional neural network and achieves relatively high segmentation accuracy. Experimental results suggest that this method lays the groundwork for accurately measuring pig-body temperature.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Minjuan Wang, Zhen Zhong, Wanlin Gao
PROCEEDINGS OF THE THIRD INTERNATIONAL CONFERENCE ON COMPUTER SCIENCE AND APPLICATION ENGINEERING (CSAE2019)
(2019)
Proceedings Paper
Computer Science, Theory & Methods
Minjuan Wang, Zhen Zhong, Wanlin Gao
PROCEEDINGS OF THE 2ND INTERNATIONAL CONFERENCE ON COMPUTER SCIENCE AND APPLICATION ENGINEERING (CSAE2018)
(2018)
Proceedings Paper
Computer Science, Artificial Intelligence
Yihua Shi, Zheng Zhong, Jinfeng Yang
BIOMETRIC RECOGNITION
(2016)
Article
Computer Science, Information Systems
Jinfeng Yang, Zhen Zhong, Guimin Jia, Yanan Li
JOURNAL OF ELECTRICAL AND COMPUTER ENGINEERING
(2016)
Proceedings Paper
Computer Science, Artificial Intelligence
Zhen Zhong, Guimin Jia, Yihua Shi, Jinfeng Yang
BIOMETRIC RECOGNITION, CCBR 2015
(2015)
