Article
Computer Science, Artificial Intelligence
Yong Soon Tan, Kian Ming Lim, Chin Poo Lee
Summary: Hand gesture recognition is essential for human communication and interaction, and has applications in human-computer interaction and bridging language barriers. Hand-crafted and deep learning approaches can tackle challenges in vision-based hand gesture recognition, with the former focusing on specific challenges and the latter adapting to various challenges through supervised learning.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Information Systems
M. A. H. Akhand, Shuvendu Roy, Nazmul Siddique, Md Abdus Samad Kamal, Tetsuya Shimamura
Summary: This study proposes a highly accurate facial emotion recognition system based on a very deep CNN model, utilizing transfer learning technique for system construction and optimization. The method shows remarkable accuracy and superiority on two different facial image datasets, demonstrating proficiency in handling diverse profile views and frontal views.
Article
Biochemical Research Methods
Yuan Li, Xu Shi, Liping Yang, Chunyu Pu, Qijuan Tan, Zhengchun Yang, Hong Huang
Summary: This paper proposes a multi-layer collaborative generative adversarial transformer (MC-GAT) for cholangiocarcinoma (CCA) classification from hyperspectral pathological images. MC-GAT consists of a generator and a discriminator, which improve the model's generalization and discriminating power. Experimental results show that MC-GAT achieves better classification results compared to other methods.
BIOMEDICAL OPTICS EXPRESS
(2022)
Article
Materials Science, Multidisciplinary
Fadi Aldakheel, Celal Soyarslan, Hari Subramani Palanisamy, Elsayed Saber Elsayed
Summary: Computational material modeling using CNN provides a solution for efficient and accurate modeling in heterogeneous materials, reducing the development costs and speeding up the design process.
MECHANICS OF MATERIALS
(2023)
Article
Chemistry, Analytical
Xiao Zhou, Yuanhang Mao, Miao Gu, Zhen Cheng
Summary: Researchers developed a microfluidic system embedded with a weakly supervised cell counting network (WSCNet) to generate microfluidic droplets, evaluate their quality, and accurately recognize the locations of encapsulated cells.
Article
Computer Science, Information Systems
Leila Boussaad, Aldjia Boucetta
Summary: This paper examines the effectiveness of deep-learning based methods for age-invariant face recognition. Five popular pre-trained deep-convolutional neural network models are evaluated on a widely used face-aging database, and the AlexNet model is found to be the most promising for feature extraction. The experimental results demonstrate the potential of convolutional neural networks in face recognition across age progression.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Environmental Sciences
Haoyu Zhang, Lei Yu, Yushi Chen, Yinsheng Wei
Summary: In this study, a complex-valued CNN was utilized to recognize radar jamming signals, showing better recognition accuracy compared to traditional real-valued CNNs. To speed up recognition time, a fast CV-CNN based on pruning was proposed. Experimental results demonstrated good performance in terms of accuracy and speed for these methods, indicating the potential of CV-CNN-based approaches in radar signal processing.
Article
Engineering, Marine
Zhuoyi Li, Deshan Chen, Tsz Leung Yip, Jinfen Zhang
Summary: This paper proposes a novel target recognition method for side scan sonar (SSS) images in varied underwater environment, called YOLO-slimming, based on convolutional neural network (CNN). The method introduces efficient feature encoders to strengthen the representation of feature maps. Channel-level sparsity regularization in model training is performed to speed up the inference performance. By using a sonar image simulation method based on deep style transfer (ST), the scarcity of SSS images is overcome. The results on the SSS image dataset show that the method can reduce calculations and improve the inference speed with a mean average precision (mAP) of 95.3 and at least 45 frames per second (FPS) on an embedded GPU, proving its feasibility in practical application and the potential to formulate an image-based real-time underwater target recognition system.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2023)
Article
Chemistry, Analytical
Zihao Jin, Zhiming Xing, Yiran Wang, Shuqi Fang, Xiumin Gao, Xiangmei Dong
Summary: In recent years, there has been a growing interest in emotion recognition research based on cerebral blood oxygen signals. This study utilized a functional near-infrared spectroscopy system to record brain signals while participants watched videos, and proposed a CNN-Transformer network to improve the classification accuracy of ternary emotions. Experimental results showed that the CNN-Transformer network achieved higher accuracy compared to CNN.
Article
Chemistry, Analytical
Ling-Feng Shi, Zhong-Ye Liu, Ke-Jun Zhou, Yifan Shi, Xiao Jing
Summary: This paper proposes a method called SConvLSTM for gait recognition using multimodal wearable inertial sensors. By employing 1D-CNN and bidirectional LSTM network, the method can automatically extract features from raw acceleration and gyroscope signals, while retaining the time-series characteristics of the data and compressing feature vector length. Experimental results demonstrate that SConvLSTM outperforms most of the current best methods on three benchmark datasets.
Article
Computer Science, Artificial Intelligence
Hongmin Gao, Hongyi Wu, Zhonghao Chen, Yunfei Zhang, Yiyan Zhang, Chenming Li
Summary: This paper proposes a novel multiscale spectral-spatial cross-extraction network (MSSCEN) for hyperspectral image classification, achieving the best classification accuracy in experiments. By introducing spectral-spatial features cross extraction module (SSCEM) and an independent data augmentation module, the method fully utilizes the changes caused by convolution.
IET IMAGE PROCESSING
(2022)
Article
Computer Science, Artificial Intelligence
Kaixuan Yao, Feilong Cao, Yee Leung, Jiye Liang
Summary: This paper proposes a method to compress deep neural networks based on interpretability, achieving effective compression while providing a better interpretation of the deep learning process. By utilizing single-layer filter pruning, the entire DNN model can be compressed layer by layer, resulting in reduced computation cost and storage requirements for implementing complex DNN models in small mobile devices.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Artificial Intelligence
Ziying Yang, Wenyan He, Xijian Fan, Tardi Tjahjadi
Summary: In this paper, a plant recognition network called PlantNet is proposed, which is based on transfer learning and bilinear convolutional neural network. The network achieves high recognition accuracy in high-throughput phenotyping requirements. Experimental results show that the proposed bilinear model outperforms other methods on the Arabidopsis dataset, demonstrating good generalization ability and robust performance.
Article
Computer Science, Information Systems
Saeed Mohsen, Anas M. Ali, Ahmed Emam
Summary: Two different CNN models are proposed in this paper for the recognition of modulation techniques, and they are applied to two different datasets: RadioML2016.10a and a dataset with 24,460 images. The experimental results show that these CNN models achieve high performance in the classification of modulation techniques.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Sanghyun Choo, Hoonseok Park, Sangyeon Kim, Donghyun Park, Jae-Yoon Jung, Sangwon Lee, Chang S. Nam
Summary: This study examined the effectiveness of multi-task learning (MTL) for raw EEG-based convolutional neural networks (CNNs) in emotion recognition with auxiliary context information. The results showed that the MTL classifier had a significantly higher classification accuracy and improved the performance of single-task learnings (STLs) for both emotion and context, with ShallowConvNet being the best network architecture among the considered CNNs.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
