https://reunir.unir.net/handle/123456789/13673 Fri, 08 Nov 2024 13:43:22 GMT 2024-11-08T13:43:22Z https://reunir.unir.net/handle/123456789/13896 Editor's Note Shu, Lei; Rodrigues, Joel J.P.C.; Cohn, Anthony G.; Mao, Qirong; Li, Maozhen As the Internet of Things (IoT) further develops and expands to the Internet of Everything (IoE), high-speed multimedia streaming data processing, analysis, and shorter response times are increasingly becoming the demands of today. Driven by the Internet of Things (IoT), a new computing paradigm, Edge computing, is currently developing rapidly. Compared with traditional centralized generalpurpose computing, Edge computing is a distributed architecture. The operations of applications, data and services are moved from the central node of the network to the edge nodes on the network logic for processing. Under this structure, the analysis of data and the generation of knowledge are closer to the source of the data, so it is more suitable for processing. However, with the rapid development of 5G, IoT and other services and scenarios, there are more and more intelligent terminal devices. Multimedia streaming processing in IoT becomes a very prominent problem. To overcome this problem, the adoption of intelligent Edge or Artificial Intelligence (AI) powered Edge computing (Edge-AI) can achieve the goals of lower cost, higher security, lower latency, and ease of management. Recently, many network modeling methods, computing algorithms, and signal processing technologies have been successfully developed and applied to multimedia streaming processing in IoT with Edge Intelligence. A total of 13 papers are presented in this special issue for the purpose of collecting the latest developments and results on this research topic. We divide them into three categories: production and life applications, security, and text and image processing. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-12-13T11:02:05Z No. of bitstreams: 1 ijimai_7_5_0 (1).pdf: 57934 bytes, checksum: f5174c18dbace77d839513db23c74daa (MD5); Made available in DSpace on 2022-12-13T11:02:05Z (GMT). No. of bitstreams: 1 ijimai_7_5_0 (1).pdf: 57934 bytes, checksum: f5174c18dbace77d839513db23c74daa (MD5) https://reunir.unir.net/handle/123456789/13896 https://reunir.unir.net/handle/123456789/13713 Electromiographic Signal Processing Using Embedded Artificial Intelligence: An Adaptive Filtering Approach Proaño-Guevara, Daniel; Blanco Valencia, Xiomara Patricia; Rosero-Montalvo, Paul D.; Peluffo-Ordóñez, Diego H. In recent times, Artificial Intelligence (AI) has become ubiquitous in technological fields, mainly due to its ability to perform computations in distributed systems or the cloud. Nevertheless, for some applications -as the case of EMG signal processing- it may be highly advisable or even mandatory an on-the-edge processing, i.e., an embedded processing methodology. On the other hand, sEMG signals have been traditionally processed using LTI techniques for simplicity in computing. However, making this strong assumption leads to information loss and spurious results. Considering the current advances in silicon technology and increasing computer power, it is possible to process these biosignals with AI-based techniques correctly. This paper presents an embedded-processing-based adaptive filtering system (here termed edge AI) being an outstanding alternative in contrast to a sensor-computer- actuator system and a classical digital signal processor (DSP) device. Specifically, a PYNQ-Z1 embedded system is used. For experimental purposes, three methodologies on similar processing scenarios are compared. The results show that the edge AI methodology is superior to benchmark approaches by reducing the processing time compared to classical DSPs and general standards while maintaining the signal integrity and processing it, considering that the EMG system is not LTI. Likewise, due to the nature of the proposed architecture, handling information exhibits no leakages. Findings suggest that edge computing is suitable for EMG signal processing when an on-device analysis is required. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-24T12:32:00Z No. of bitstreams: 1 ijimai_7_5_5.pdf: 2505299 bytes, checksum: a1cd541c74e13838cb329bb21f83d7ee (MD5); Made available in DSpace on 2022-10-24T12:32:00Z (GMT). No. of bitstreams: 1 ijimai_7_5_5.pdf: 2505299 bytes, checksum: a1cd541c74e13838cb329bb21f83d7ee (MD5) https://reunir.unir.net/handle/123456789/13713 https://reunir.unir.net/handle/123456789/13712 ED-Dehaze Net: Encoder and Decoder Dehaze Network Zhang, Hongqi; Wei, Yixiong; Zhou, Hongqiao; Wu, Qianhao The presence of haze will significantly reduce the quality of images, such as resulting in lower contrast and blurry details. This paper proposes a novel end-to-end dehazing method, called Encoder and Decoder Dehaze Network (ED-Dehaze Net), which contains a Generator and a Discriminator. In particular, the Generator uses an Encoder-Decoder structure to effectively extract the texture and semantic features of hazy images. Between the Encoder and Decoder we use Multi-Scale Convolution Block (MSCB) to enhance the process of feature extraction. The proposed ED-Dehaze Net is trained by combining Adversarial Loss, Perceptual Loss and Smooth L1 Loss. Quantitative and qualitative experimental results showed that our method can obtain the state-of-the-art dehazing performance. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-24T12:19:23Z No. of bitstreams: 1 ijimai_7_5_11.pdf: 1073103 bytes, checksum: d17c5e128ba713f448c3f5ce626236bb (MD5); Made available in DSpace on 2022-10-24T12:19:23Z (GMT). No. of bitstreams: 1 ijimai_7_5_11.pdf: 1073103 bytes, checksum: d17c5e128ba713f448c3f5ce626236bb (MD5) https://reunir.unir.net/handle/123456789/13712 https://reunir.unir.net/handle/123456789/13711 Interactive Causal Correlation Space Reshape for Multi-Label Classification Zhang, Chao; Cheng, Yusheng; Wang, Yibin; Xu, Yuting Most existing multi-label classification models focus on distance metrics and feature spare strategies to extract specific features of labels. Those models use the cosine similarity to construct the label correlation matrix to constraint solution space, and then mine the latent semantic information of the label space. However, the label correlation matrix is usually directly added to the model, which ignores the interactive causality of the correlation between the labels. Considering the label-specific features based on the distance method merely may have the problem of distance measurement failure in the high-dimensional space, while based on the sparse weight matrix method may cause the problem that parameter is dependent on manual selection. Eventually, this leads to poor classifier performance. In addition, it is considered that logical labels cannot describe the importance of different labels and cannot fully express semantic information. Based on these, we propose an Interactive Causal Correlation Space Reshape for Multi-Label Classification (CCSRMC) algorithm. Firstly, the algorithm constructs the label propagation matrix using characteristic that similar instances can be linearly represented by each other. Secondly, label co-occurrence matrix is constructed by combining the conditional probability test method, which is based on the label propagation reshaping the label space to rich label semantics. Then the label co-occurrence matrix combines with the label correlation matrix to construct the label interactive causal correlation matrix to perform multi-label classification learning on the obtained numerical label matrix. Finally, the algorithm in this paper is compared with multiple advanced algorithms on multiple benchmark multi-label datasets. The results show that considering the interactive causal label correlation can reduce the redundant information in the model and improve the performance of the multi-label classifier. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-24T12:11:48Z No. of bitstreams: 1 ijimai_7_5_13.pdf: 1807426 bytes, checksum: 96ad73357576576500aa03ce6588e770 (MD5); Made available in DSpace on 2022-10-24T12:11:48Z (GMT). No. of bitstreams: 1 ijimai_7_5_13.pdf: 1807426 bytes, checksum: 96ad73357576576500aa03ce6588e770 (MD5) https://reunir.unir.net/handle/123456789/13711 https://reunir.unir.net/handle/123456789/13710 Variational Learning for the Inverted Beta-Liouville Mixture Model and Its Application to Text Categorization Ling, Yongfa; Guan, Wenbo; Ruan, Qiang; Song, Heping; Lai, Yuping he finite invert Beta-Liouville mixture model (IBLMM) has recently gained some attention due to its positive data modeling capability. Under the conventional variational inference (VI) framework, the analytically tractable solution to the optimization of the variational posterior distribution cannot be obtained, since the variational object function involves evaluation of intractable moments. With the recently proposed extended variational inference (EVI) framework, a new function is proposed to replace the original variational object function in order to avoid intractable moment computation, so that the analytically tractable solution of the IBLMM can be derived in an effective way. The good performance of the proposed approach is demonstrated by experiments with both synthesized data and a real-world application namely text categorization. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-24T11:57:46Z No. of bitstreams: 1 ijimai_7_5_9.pdf: 1568609 bytes, checksum: dbf2880f5751c6294df6ba662961d4aa (MD5); Made available in DSpace on 2022-10-24T11:57:46Z (GMT). No. of bitstreams: 1 ijimai_7_5_9.pdf: 1568609 bytes, checksum: dbf2880f5751c6294df6ba662961d4aa (MD5) https://reunir.unir.net/handle/123456789/13710 https://reunir.unir.net/handle/123456789/13709 Design of Integrated Artificial Intelligence Techniques for Video Surveillance on IoT Enabled Wireless Multimedia Sensor Networks Mansour, Romany F.; Soto, Carlos; Soto-Díaz, Roosvel; Escorcia Gutierrez, José; Gupta, Deepak; Khanna, Ashish The recent advancements in the Internet of Things (IoT) and Wireless Multimedia Sensor Networks (WMSN) made high-speed multimedia streaming, data processing, and essential analytics processes with minimal delay. Multimedia sensors used in WMSN-based surveillance applications are beneficial helpful in attaining accurate and elaborate details. However, it has become essential to design an effective and lightweight solution for data traffic management in WMSN owing to the massive quantities of data, generated by multimedia sensors. The development of Artificial Intelligence (AI) and Machine Learning (ML) techniques can be leveraged to investigate, collect, store, and process multimedia streaming data for decision-making in real-time scenarios. In this aspect, the current study develops an Integrated AI technique for Video Surveillance in IoT-enabled WMSN, called IAIVS-WMSN. The proposed IAIVS-WMSN technique aims to design a practical scheme for object detection and data transmission in WMSN. The proposed IAIVS-WMSN approach encompasses three stages: object detection, image compression, and clustering. The Mask Regional Convolutional Neural Network (Mask RCNN) technique is primarily utilized for object detection in the target region. Besides, Neighbourhood Correlation Sequence-based Image Compression (NCSIC) technique is applied to reduce data transmission. Finally, Artificial Flora Algorithm (AFA)-based clustering technique is designed for the election of Cluster Heads (CHs) and construction clusters. The design of object detection with compression and clustering techniques for WMSN shows the novelty of the work. These three processes’ designs enable one to accomplish effective data transmission in IoT-enabled WMSN. The researchers conducted multiple simulations to highlight the supreme performance of the IAIVS-WMSN approach. The simulation outcomes inferred the enhanced performance of the IAIVS-WMSN algorithm to the existing approaches. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-24T11:47:16Z No. of bitstreams: 1 ijimai_7_5_2.pdf: 945311 bytes, checksum: 73bae6ed6679f9f2d0cda8b9026fa47d (MD5); Made available in DSpace on 2022-10-24T11:47:16Z (GMT). No. of bitstreams: 1 ijimai_7_5_2.pdf: 945311 bytes, checksum: 73bae6ed6679f9f2d0cda8b9026fa47d (MD5) https://reunir.unir.net/handle/123456789/13709 https://reunir.unir.net/handle/123456789/13708 Content-Based Hyperspectral Image Compression Using a Multi-Depth Weighted Map With Dynamic Receptive Field Convolution Pan, Shaoming; Gu, XiaoLin; Chong, Yanwen; Guo, Yuanyuan In content-based image compression, the importance map guides the bit allocation based on its ability to represent the importance of image contents. In this paper, we improve the representational power of importance map using Squeeze-and-Excitation (SE) block, and propose multi-depth structure to reconstruct non-important channel information at low bit rates. Furthermore, Dynamic Receptive Field convolution (DRFc) is introduced to improve the ability of normal convolution to extract edge information, so as to increase the weight of edge content in the importance map and improve the reconstruction quality of edge regions. Results indicate that our proposed method can extract an importance map with clear edges and fewer artifacts so as to provide obvious advantages for bit rate allocation in content-based image compression. Compared with typical compression methods, our proposed method can greatly improve the performance of Peak Signal-to-Noise Ratio (PSNR), structural similarity (SSIM) and spectral angle (SAM) on three public datasets, and can produce a much better visual result with sharp edges and fewer artifacts. As a result, our proposed method reduces the SAM by 42.8% compared to the recently SOTA method to achieve the same low bpp (0.25) on the KAIST dataset. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-24T11:13:28Z No. of bitstreams: 1 ijimai_7_5_10.pdf: 2184511 bytes, checksum: 5372b03215f7ec31be4ee91a33d8737f (MD5); Made available in DSpace on 2022-10-24T11:13:28Z (GMT). No. of bitstreams: 1 ijimai_7_5_10.pdf: 2184511 bytes, checksum: 5372b03215f7ec31be4ee91a33d8737f (MD5) https://reunir.unir.net/handle/123456789/13708 https://reunir.unir.net/handle/123456789/13682 Improvement in Quality of Service Against Doppelganger Attacks for Connected Network Choudhary, Deepak; Pahuja, Roop Because they are in a high-risk location, remote sensors are vulnerable to malicious ambushes. A doppelganger attack, in which a malicious hub impersonates a legitimate network junction and then attempts to take control of the entire network, is one of the deadliest types of ambushes. Because remote sensor networks are portable, hub doppelganger ambushes are particularly ineffective in astute wellness contexts. Keeping the framework safe from hostile hubs is critical because the information in intelligent health frameworks is so sensitive. This paper developed a new Steering Convention for Vitality Effective Systems (SC-VFS) technique for detecting doppelganger attacks in IoT-based intelligent health applications such as a green corridor for transplant pushback. This method's main advantage is that it improves vitality proficiency, a critical constraint in WSN frameworks. To emphasize the suggested scheme's execution, latency, remaining vitality, throughput, vitality effectiveness, and blunder rate are all used. To see how proper the underutilized technique is compared to the existing Half Breed Multi-Level Clustering (HMLC) computation. The suggested approach yields latency of 0.63ms and 0.6ms, respectively, when using dead hubs and keeping a strategic distance from doppelganger assault. Furthermore, during the 2500 cycles, the suggested system achieves the highest remaining vitality of 49.5J. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-20T11:18:26Z No. of bitstreams: 1 ijimai_7_5_6.pdf: 1342027 bytes, checksum: 7c92439ec000450b620e83a1e9180138 (MD5); Made available in DSpace on 2022-10-20T11:18:26Z (GMT). No. of bitstreams: 1 ijimai_7_5_6.pdf: 1342027 bytes, checksum: 7c92439ec000450b620e83a1e9180138 (MD5) https://reunir.unir.net/handle/123456789/13682 https://reunir.unir.net/handle/123456789/13681 A Security Situation Awareness Approach for IoT Software Chain Based on Markov Game Model Zhu, Xudong; Deng, Honggao Since Internet of Things (IoT) has been widely used in our daily life nowadays, it is regarded as a promising and popular application of the Internet, and has attracted more and more attention. However, IoT is also suffered by some security problems which seriously affect the implementation of IoT system. Similar to traditional software, IoT software is always threated by many vulnerabilities, thus how to evaluate the security situation of IoT software chain becomes a basic requirement. In this paper, A framework of security situation awareness for IoT software chain is proposed, which mainly includes two processes: IoT security situation classification based on support vector machine and security situation awareness based on Markov game model. The proposed method firstly constructs a classification model using support vector machine (IoT) to automatically evaluates the security situation of IoT software chain. Based on the situation classification, we further proposed to adopt Markov model to simulate and predict the next behaviors of participants that involved in IoT system. Additionally, we have designed and developed a security situation awareness system for IoT software chain, the developed system supports the detection of typical IoT vulnerabilities and inherits more than 20 vulnerability detection methods, which shows great potential in IoT system protection. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-20T11:14:03Z No. of bitstreams: 1 ijimai_7_5_7.pdf: 448210 bytes, checksum: a3a6c4eb9dbbf54f4b39f20b17f5037a (MD5); Made available in DSpace on 2022-10-20T11:14:03Z (GMT). No. of bitstreams: 1 ijimai_7_5_7.pdf: 448210 bytes, checksum: a3a6c4eb9dbbf54f4b39f20b17f5037a (MD5) https://reunir.unir.net/handle/123456789/13681 https://reunir.unir.net/handle/123456789/13680 A Diverse Domain Generative Adversarial Network for Style Transfer on Face Photographs Tahir, Rabia; Cheng, Keyang; Memon, Bilal Ahmed; Liu, Qing The applications of style transfer on real time photographs are very trending now. This is used in various applications especially in social networking sites such as SnapChat and beauty cameras. A number of style transfer algorithms have been proposed but they are computationally expensive and generate artifacts in output image. Besides, most of research work only focuses on some traditional painting style transfer on real photographs. However, our work is unique as it considers diverse style domains to be transferred on real photographs by using one model. In this paper, we propose a Diverse Domain Generative Adversarial Network (DD-GAN) which performs fast diverse domain style translation on human face images. Our work is highly efficient and focused on applying different attractive and unique painting styles to human photographs while keeping the content preserved after translation. Moreover, we adopt a new loss function in our model and use PReLU activation function which improves and fastens the training procedure and helps in achieving high accuracy rates. Our loss function helps the proposed model in achieving better reconstructed images. The proposed model also occupies less memory space during training. We use various evaluation parameters to inspect the accuracy of our model. The experimental results demonstrate the effectiveness of our method as compared to state-of-the-art results. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-20T11:09:34Z No. of bitstreams: 1 ijimai_7_5_12.pdf: 1419197 bytes, checksum: f6514d66ec1ea8bc2bb0aaa16de27c0e (MD5); Made available in DSpace on 2022-10-20T11:09:35Z (GMT). No. of bitstreams: 1 ijimai_7_5_12.pdf: 1419197 bytes, checksum: f6514d66ec1ea8bc2bb0aaa16de27c0e (MD5) https://reunir.unir.net/handle/123456789/13680 https://reunir.unir.net/handle/123456789/13677 STAIBT: Blockchain and CP-ABE Empowered Secure and Trusted Agricultural IoT Blockchain Terminal Zhang, Guofeng; Chen, Xiao; Zhang, Lei; Feng, Bin; Guo, Xuchao; Liang, Jingyun; Zhang, Yanan The integration of agricultural Internet of Things (IoT) and blockchain has become the key technology of precision agriculture. How to protect data privacy and security from data source is one of the difficult issues in agricultural IoT research. This work integrates cryptography, blockchain and Interplanetary File System (IPFS) technologies, and proposes a general IoT blockchain terminal system architecture, which strongly supports the integration of the IoT and blockchain technology. This research innovatively designed a fine-grained and flexible terminal data access control scheme based on the ciphertext-policy attribute-based encryption (CP-ABE) algorithm. Based on CP-ABE and DES algorithms, a hybrid data encryption scheme is designed to realize 1-to-N encrypted data sharing. A "horizontal + vertical" IoT data segmentation scheme under blockchain technology is proposed to realize the classified release of different types of data on the blockchain. The experimental results show that the design scheme can ensure data access control security, privacy data confidentiality, and data high-availability security. This solution significantly reduces the complexity of key management, can realize efficient sharing of encrypted data, flexibly set access control strategies, and has the ability to store large data files in the agricultural IoT. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-19T13:36:18Z No. of bitstreams: 1 ijimai_7_5_8.pdf: 1468614 bytes, checksum: f7a8958add3d998f5b782372ef4574b2 (MD5); Made available in DSpace on 2022-10-19T13:36:18Z (GMT). No. of bitstreams: 1 ijimai_7_5_8.pdf: 1468614 bytes, checksum: f7a8958add3d998f5b782372ef4574b2 (MD5) https://reunir.unir.net/handle/123456789/13677 https://reunir.unir.net/handle/123456789/13676 A Method of the Coverage Ratio of Street Trees Based on Deep Learning Han, Wen; Cao, Lei; Xu, Sheng The street trees coverage ratio provides reliable data support for urban ecological environment assessment, which plays an important part in the ecological environment index calculation. Aiming at the statistical estimation of urban street trees coverage ratio, an integrated model based on YOLOv4 and Unet network for detecting and extracting street trees from remote sensing images is proposed, and obtain the estimated street trees coverage ratio in images accurately. The experiments are carried out under self-made dataset, and the results show that the accuracy of street trees detection is 94.91%, and the street trees coverage ratio is 16.30% and 13.81% in the two experimental urban scenes. The MIoU of contour extraction is 98.25%, and the estimated coverage accuracy is improved by 6.89% and 5.79%, respectively. The result indicates that the proposed model achieves the automation of contour extraction of street trees and more accurate estimation of street trees coverage ratio. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-19T13:29:33Z No. of bitstreams: 1 ijimai_7_5_3.pdf: 1317222 bytes, checksum: c7affc1942dda0dbd6e289c687e80f75 (MD5); Made available in DSpace on 2022-10-19T13:29:33Z (GMT). No. of bitstreams: 1 ijimai_7_5_3.pdf: 1317222 bytes, checksum: c7affc1942dda0dbd6e289c687e80f75 (MD5) https://reunir.unir.net/handle/123456789/13676 https://reunir.unir.net/handle/123456789/13675 Improved GWO Algorithm for UAV Path Planning on Crop Pest Monitoring Ding, Qun; Xu, Xiaolong Agricultural information monitoring is the monitoring of the agricultural production process, and its task is to monitor the growth process of major crops systematically. When assessing the pest situation of crops in this process, the traditional satellite monitoring method has the defects of poor real-time and high operating cost, whereas the pest monitoring through Unmanned Aerial Vehicles (UAVs) effectively solves the above problems, so this method is widely used. An important key issue involved in monitoring technology is path planning. In this paper, we proposed an Improved Grey Wolf Optimization algorithm, IGWO, to realize the flight path planning of UAV in crop pest monitoring. A map environment model is simulated, and information traversal is performed, then the search of feasible paths for UAV flight is carried out by the Grey Wolf Optimization algorithm (GWO). However, the algorithm search process has the defect of falling into local optimum which leading to path planning failure. To avoid such a situation, we introduced the probabilistic leap mechanism of the Simulated Annealing algorithm (SA). Besides, the convergence factor is modified with an exponential decay mode for improving the convergence rate of the algorithm. Compared with the GWO algorithm, IGWO has the 8.3%, 16.7%, 28.6% and 39.6% lower total cost of path distance on map models with precision of 15, 20, 25 and 30 respectively, and also has better path planning results in contrast to other swarm intelligence algorithms. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-19T13:19:20Z No. of bitstreams: 1 ijimai_7_5_4.pdf: 1625214 bytes, checksum: f64c3c28f5c8645d6ca6a10ead4fcc1e (MD5); Made available in DSpace on 2022-10-19T13:19:20Z (GMT). No. of bitstreams: 1 ijimai_7_5_4.pdf: 1625214 bytes, checksum: f64c3c28f5c8645d6ca6a10ead4fcc1e (MD5) https://reunir.unir.net/handle/123456789/13675 https://reunir.unir.net/handle/123456789/13674 An EEG Signal Recognition Algorithm During Epileptic Seizure Based on Distributed Edge Computing Qiu, Shi; Cheng, Keyang; Zhou, Tao; Tahir, Rabia; Ting, Liang Epilepsy is one kind of brain diseases, and its sudden unpredictability is the main cause of disability and even death. Thus, it is of great significance to identify electroencephalogram (EEG) during the seizure quickly and accurately. With the rise of cloud computing and edge computing, the interface between local detection and cloud recognition is established, which promotes the development of portable EEG detection and diagnosis. Thus, we construct a framework for identifying EEG signals in epileptic seizure based on cloud-edge computing. The EEG signals are obtained in real time locally, and the horizontal viewable model is established at the edge to enhance the internal correlation of the signals. The Takagi-Sugeno-Kang (TSK) fuzzy system is established to analyze the epileptic signals. In the cloud, the fusion of clinical features and signal features is established to establish a deep learning framework. Through local signal acquisition, edge signal processing and cloud signal recognition, the diagnosis of epilepsy is realized, which can provide a new idea for the real-time diagnosis and feedback of EEG during epileptic seizure. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-10-19T13:13:42Z No. of bitstreams: 1 ijimai_7_5_1.pdf: 1175870 bytes, checksum: e40110c06e611ff697052c4273db5617 (MD5); Made available in DSpace on 2022-10-19T13:13:42Z (GMT). No. of bitstreams: 1 ijimai_7_5_1.pdf: 1175870 bytes, checksum: e40110c06e611ff697052c4273db5617 (MD5) https://reunir.unir.net/handle/123456789/13674