International Journal of Intelligent Systems最新文献

With the rapid development of media, the role of video quality assessment (VQA) is becoming increasingly significant. VQA has applications in many domains. For example, in the field of remote medical diagnosis, it can enhance the quality of video communication between doctors and patients. Besides, in sports broadcasting, it can improve video clarity. Within VQA, the human visual system (HVS) is a crucial component that should be taken into consideration. Considering that attention is guided by goal-driven and top-down factors, such as anticipated locations or some attractive frames within the video, we propose a blind VQA algorithm based on spatial-temporal attention model. Specifically, we first use two pretrained convolutional networks to extract low-level static-dynamic fusion features. Then, a spatial attention-guided model is established to get more representative features of frame-level quality perception. Next, through a temporal attention-guided model, the video-level features are obtained. Finally, the features are fed into a regression model to calculate the final video quality score. The experiments conducted on seven VQA databases reach the state-of-the-art performance, demonstrating the effectiveness of our proposed method.

Lightweight, high-performance networks are important in vision perception systems. Recent research on convolutional neural networks has shown that attention mechanisms can significantly improve the network performance. However, existing approaches either ignore the significance of using both types of attention mechanisms (channel and space) simultaneously or increase the model complexity. In this study, we propose the adaptive attention module (AAM), which is a truly lightweight yet effective module that comprises channel and spatial submodules to balance model performance and complexity. The AAM initially utilizes the channel submodule to generate intermediate channel-refined features. In this module, an adaptive mechanism enables the model to autonomously learn the weights between features extracted by global max pooling and global average pooling to adapt to different stages of the model, thus enhancing performance. The spatial submodule employs a group-interact-aggregate strategy to enhance the expression of important features. It groups the intermediate channel-refined features along the channel dimension into multiple subfeatures for parallel processing and generates spatial attention feature descriptors and channelwise refined subfeatures for each subfeature; subsequently, it aggregates all the refined subfeatures and employs a “channel shuffle” operator to transfer information between different subfeatures, thereby generating the final refined features and adaptively emphasizing important regions. Additionally, AAM is a plug-and-play architectural unit that can be directly used to replace standard convolutions in various convolutional neural networks. Extensive tests on CIFAR-100, ImageNet-1k, BDD100K, and MS COCO demonstrate that AAM improves the baseline network performance under various models and tasks, thereby validating its versatility.

This article proposes a novel Mahjong game model, LsAc ^∗-MJ, designed to address challenges posed by data scarcity, difficulty in leveraging contextual information, and the computational resource-intensive nature of self-play zero-shot learning. The model is applied to Japanese Mahjong for experiments. LsAc ^∗-MJ employs long short-term memory (LSTM) neural networks, utilizing hidden nodes to store and propagate contextual historical information, thereby enhancing decision accuracy. Additionally, the paper introduces an optimized Advantage Actor-Critic (A2C) algorithm incorporating an experience replay mechanism to enhance the model’s decision-making capabilities and mitigate convergence difficulties arising from strong data correlations. Furthermore, the paper presents a two-stage training approach for self-play deep reinforcement learning models guided by expert knowledge, thereby improving training efficiency. Extensive ablation experiments and performance comparisons demonstrate that, in contrast to other typical deep reinforcement learning models on the RLcard platform, the LsAc ^∗-MJ model consumes lower computational and time resources, has higher training efficiency, faster average decision time, higher win-rate, and stronger decision-making ability.

Computational offloading in mobile edge computing (MEC) systems provides an efficient solution for resource-intensive applications on devices. However, the frequent communication between devices and edge servers increases the traffic within the network, thereby hindering significant improvements in latency. Furthermore, the benefits of MEC cannot be fully realized when the communication link utilized for offloading tasks experiences severe attenuation. Fortunately, reconfigurable intelligent surfaces (RISs) can mitigate propagation-induced impairments by adjusting the phase shifts imposed on the incident signals using their passive reflecting elements. This paper investigates the performance gains achieved by deploying multiple RISs in MEC systems under energy-constrained conditions to minimize the overall system latency. Considering the high coupling among variables such as the selection of multiple RISs, optimization of their phase shifts, transmit power, and MEC offloading volume, the problem is formulated as a nonconvex problem. We propose two approaches to address this problem. First, we employ an alternating optimization approach based on semidefinite relaxation (AO-SDR) to decompose the original problem into two subproblems, enabling the alternating optimization of multi-RIS communication and MEC offloading volume. Second, due to its capability to model and learn the optimal phase adjustment strategies adaptively in dynamic and uncertain environments, deep reinforcement learning (DRL) offers a promising approach to enhance the performance of phase optimization strategies. We leverage DRL to address the joint design of MEC-offloading volume and multi-RIS communication. Extensive simulations and numerical analysis results demonstrate that compared to conventional MEC systems without RIS assistance, the multi-RIS-assisted schemes based on the AO-SDR and DRL methods achieve a reduction in latency by 23.5% and 29.6%, respectively.

In this paper, a novel integrated long short-term memory (LSTM) network and dynamic update model are proposed for long-term object tracking in video images. The LSTM network tracking method is introduced to improve the effect of tracking failure caused by target occlusion. Stable tracking of the target is achieved using the LSTM method to predict the motion trajectory of the target when it is occluded and dynamically updating the tracking template. First, in target tracking, global average peak-to-correlation energy (GAPCE) is used to determine whether the tracking target is blocked or temporarily disappearing such that the follow-up response tracking strategy can be adjusted accordingly. Second, the data with target motion characteristics are utilized to train the designed LSTM model to obtain an offline model, which effectively predicts the motion trajectory during the period when the target is occluded or has disappeared. Therefore, it can be captured again when the target reappears. Finally, in the dynamic template adjustment stage, the historical information of the target movement is combined, and the corresponding value of the current target is compared with the historical response value to realize the dynamic adjustment of the target tracking template. Compared with the current mainstream efficient convolution operators, namely, the E.T.Track, ToMP, KeepTrack, and RTS algorithms, on the OTB100 and LaSOT datasets, the proposed algorithm increases the distance precision by 9.9% when the distance threshold is 5 pixels, increases the overlap success rate by 0.94% when the overlap threshold is 0.75, and decreases the center location error by 18.9%. The proposed method has higher tracking accuracy and robustness and is more suitable for long-term tracking of targets in actual scenarios than are the main approaches.

Brain tumors significantly impact human health due to their complexity and the challenges in early detection and treatment. Accurate diagnosis is crucial for effective intervention, but existing methods often suffer from limitations in accuracy and efficiency. To address these challenges, this study presents a novel deep learning (DL) approach utilizing the EfficientNet family for enhanced brain tumor classification and detection. Leveraging a comprehensive dataset of 3064 T1-weighted CE MRI images, our methodology incorporates advanced preprocessing and augmentation techniques to optimize model performance. The experiments demonstrate that EfficientNetB(07) achieved 99.14%, 98.76%, 99.07%, 99.69%, 99.07%, 98.76%, 98.76%, and 99.07% accuracy, respectively. The pinnacle of our research is the EfficientNetB3 model, which demonstrated exceptional performance with an accuracy rate of 99.69%. This performance surpasses many existing state-of-the-art (SOTA) techniques, underscoring the efficacy of our approach. The precision of our high-accuracy DL model promises to improve diagnostic reliability and speed in clinical settings, facilitating earlier and more effective treatment strategies. Our findings suggest significant potential for improving patient outcomes in brain tumor diagnosis.

The rapid development of vehicle-to-vehicle (V2V) communication technology provides more opportunities to improve traffic safety and efficiency, which facilitates the exchange of multivehicle information to mine potential patterns and hidden associations in vehicle trajectory prediction. To address the importance of fine-grained vehicle interaction modelling in vehicle trajectory prediction, the present work proposes an integrated vehicle trajectory prediction model that combines the multihead attention mechanism with long short-term memory (IMA-LSTM) in multivehicle interaction scenarios. Compared to existing studies, a dedicated feature extraction module including both individual features and interactive features is designed and sophisticated multihead attention mechanism is applied with LSTM framework to capture the variation of spatial-temporal interactions between vehicles. The performance of the proposed model in different scenarios is examined using both the high-D and the NGSIM dataset through comprehensive comparison experiments. The results indicate that the proposed IMA-LSTM model presents great improvement in vehicle trajectory prediction performance in different scenarios compared to the model that does not consider multivehicle interaction features. Moreover, it outperforms other existing models in 3–5s prediction horizons and such superiority is more evident in left lane-changing (LLC) scenarios than lane-keeping (LK) and right lane-changing (RLC) scenarios. The outcomes fully address the importance of fine-grained interactive feature modelling in improving vehicle trajectory performance in complex multivehicle interaction scenarios and could further contribute to more refined traffic safety and traffic efficiency management.

Electroencephalography (EEG) data serve as a reliable method for fatigue detection due to their intuitive representation of drivers’ mental processes. However, existing research on feature generation has overlooked the effective and automated aspects of this process. The challenge of extracting features from unpredictable and complex EEG signals has led to the frequent use of deep learning models for signal classification. Unfortunately, these models often neglect generalizability to novel subjects. To address these concerns, this study proposes the utilization of a modified deep convolutional neural network, specifically the Inception-dilated ResNet architecture. Trained on spectrograms derived from segmented EEG data, the network undergoes analysis in both temporal and spatial-frequency dimensions. The primary focus is on accurately detecting and classifying fatigue. The inherent variability of EEG signals between individuals, coupled with limited samples during fatigue states, presents challenges in fatigue detection through brain signals. Therefore, a detailed structural analysis of fatigue episodes is crucial. Experimental results demonstrate the proposed methodology’s ability to distinguish between alertness and sleepiness, achieving average accuracy rates of 98.87% and 82.73% on Figshare and SEED-VIG datasets, respectively, surpassing contemporary methodologies. Additionally, the study examines frequency bands’ relative significance to further explore participants’ inclinations in states of alertness and fatigue. This research paves the way for deeper exploration into the underlying factors contributing to mental fatigue.

The liver is the largest organ of the human body with more than 500 vital functions. In recent decades, a large number of liver patients have been reported with diseases such as cirrhosis, fibrosis, or other liver disorders. There is a need for effective, early, and accurate identification of individuals suffering from such disease so that the person may recover before the disease spreads and becomes fatal. For this, applications of machine learning are playing a significant role. Despite the advancements, existing systems remain inconsistent in performance due to limited feature selection and data imbalance. In this article, we reviewed 58 articles extracted from 5 different electronic repositories published from January 2015 to 2023. After a systematic and protocol-based review, we answered 6 research questions about machine learning algorithms. The identification of effective feature selection techniques, data imbalance management techniques, accurate machine learning algorithms, a list of available data sets with their URLs and characteristics, and feature importance based on usage has been identified for diagnosing liver disease. The reason to select this research question is, in any machine learning framework, the role of dimensionality reduction, data imbalance management, machine learning algorithm with its accuracy, and data itself is very significant. Based on the conducted review, a framework, machine learning-based liver disease diagnosis (MaLLiDD), has been proposed and validated using three datasets. The proposed framework classified liver disorders with 99.56%, 76.56%, and 76.11% accuracy. In conclusion, this article addressed six research questions by identifying effective feature selection techniques, data imbalance management techniques, algorithms, datasets, and feature importance based on usage. It also demonstrated a high accuracy with the framework for early diagnosis, marking a significant advancement.

Recently, live streaming technology has been widely utilized in areas such as online gaming, e-healthcare, and video conferencing. The increasing network and computational resources required for live streaming increase the cost of content providers and Internet Service Providers (ISPs), which may lead to increased latency or even unavailability of live streaming services. The current research primarily focuses on providing high-quality services by assessing the resource status of network nodes individually. However, the role assignment within nodes and the interconnectivity among nodes are often overlooked. To fill this gap, we propose a hierarchical game theory-based live video transmission framework to coordinate the heterogeneity of live tasks and nodes and to improve the resource utilization of nodes and the service satisfaction of users. Secondly, the service node roles are set as producers who are closer to the live streaming source and provide content, consumers who are closer to the end users and process data, and silent nodes who do not participate in the service process, and a non-cooperative game-based role competition algorithm is designed to improve the node resource utilization. Furthermore, a matching-based optimal path algorithm for media services is designed to establish optimal matching associations among service nodes to optimize the service experience. Finally, extensive simulation experiments show that our approach performs better in terms of service latency and bandwidth.