In recent years, great progress has been made in forecasting human motion in crowded scenes. However, current methods are far from practical applications due to the unbearable high computation costs, especially for en...
详细信息
ISBN:
(数字)9798350377705
ISBN:
(纸本)9798350377712
In recent years, great progress has been made in forecasting human motion in crowded scenes. However, current methods are far from practical applications due to the unbearable high computation costs, especially for encoding scene context. In addition, neglecting the partially detected trajectories makes the predicted outcome deviate from the real trajectory distribution. To handle the aforementioned concerns, we propose to represent the scene context and partially observed trajectories with sparse graphs. Customized for this special data structure, we design a hierarchical Graph Transformer Network model SparseGTN to predict multiple possible future trajectories of the target pedestrian by digesting the sparsely represented inputs. Our approach exhibits superiority over the state-of-the-art (SOTA) methods, utilizing a mere 3.42% of the number of floating point operations (FLOPs) and 0.53% of the number of model parameters. The code will be available online
⋆
.
This comprehensive review presents the role of artificial intelligence (AI), machine learning (ML), and deep learning (DL) in neuro-oncology, focussing their groundbreaking impact on the field. It gives a historical o...
详细信息
In the last few decades, Feature selection is one of the most challenging and open problem to researchers. The rapid progress in computational techniques causes the generation and recording of data in huge size. Thoug...
详细信息
ISBN:
(纸本)9781665439206
In the last few decades, Feature selection is one of the most challenging and open problem to researchers. The rapid progress in computational techniques causes the generation and recording of data in huge size. Though there exists various feature ranking methods, the processing of data is still a challenging task due to its computational complexity. The filter method has many advantages over the wrapper method. The filter methods are classifier independent and have better computational efficiency. Here, a subset of features is selected based on a certain goal function. Most of these goal functions employs the principle of information theory. Most of the algorithms in earlier studies addressed two factors, that is, maximization of relevancy and minimization of redundancy without considering the interaction among the features. This paper developed a new forward filter feature selection algorithm based on mutual information known as Maximum Dual Interaction and Maximum Feature Relevance(MDIMFR). This method considers all the three factors: relevance, redundancy, and feature interaction. This method is experimented on three datasets and compares the performance with existing methods. The results show that MDIMFR outperforms the existing competitive feature selection methods of recent studies: mRMR, JMIM and CMIM. MDIMFR also achieves good stability in average classification accuracy for a certain number of features, say k and above. Hence, these k features can be considered as an optimal feature set.
Recent works have revealed an essential paradigm in designing loss functions that differentiate individual losses vs. aggregate losses. The individual loss measures the quality of the model on a sample, while the aggr...
详细信息
The modern technology network plays an important role. Artificial intelligence to provide a higher solution for network security. As 7G networks emerge, virtual therapy platforms have begun to take center stage, offer...
详细信息
The modern technology network plays an important role. Artificial intelligence to provide a higher solution for network security. As 7G networks emerge, virtual therapy platforms have begun to take center stage, offering fast, real-time interaction and high-definition immersive experiences. While the convergence of security in health care and mobility is in many ways even more integrated into our lives, it is fraught with cybersecurity risks. This article proposes an AI-based cybersecurity framework for a 7 G-based virtual therapy platform. The framework employs advanced machine learning (ML) algorithms, predictive analytics, and adaptive threat intelligence to protect the confidentiality, integrity, and availability of data. The article covers the core building blocks like anomaly detection, secure 7 G-based virtual therapy platform authentication protocols, and privacy-preserving techniques. AI-GN gives an overview of attack mitigation using AI with a comparative analysis of existing 5G/6G frameworks and solutions and their unique aspects needed for 7G environments. This article mainly focuses on threat detection using a deep learning method, also it aims to provide a continuous authentication process. The proposed model response time is decreased by nearly 30 percent, at the same time detection rate is increased to 98.5 percent in a 7G network.
With the release of open source datasets such as nuPlan and Argoverse, the research around learning-based planners has spread a lot in the last years. Existing systems have shown excellent capabilities in imitating th...
详细信息
ISBN:
(数字)9798331505929
ISBN:
(纸本)9798331505936
With the release of open source datasets such as nuPlan and Argoverse, the research around learning-based planners has spread a lot in the last years. Existing systems have shown excellent capabilities in imitating the human driver behaviour, but they struggle to guarantee safe closed-loop driving. Conversely, optimization-based planners offer greater security in short-term planning scenarios. To confront this challenge, in this paper we propose a novel hybrid motion planner that integrates both learning-based and optimization-based techniques. Initially, a multilayer perceptron (MLP) generates a human-like trajectory, which is then refined by an optimization-based component. This component not only mini-mizes tracking errors but also computes a trajectory that is both kinematically feasible and collision-free with obstacles and road boundaries. Our model effectively balances safety and human-likeness, mitigating the trade-off inherent in these objectives. We validate our approach through simulation experiments and further demonstrate its efficacy by deploying it in real-world self-driving vehicles.
Sleep is one of the elements most vital to human life. However, the modern lifestyle continues to push people to neglect this critical requirement. With a vast majority of people falling victim to various sleep disord...
详细信息
Sleep is one of the elements most vital to human life. However, the modern lifestyle continues to push people to neglect this critical requirement. With a vast majority of people falling victim to various sleep disorders, it has become increasingly essential to have a robust system for diagnosing and treating such ailments. Sleep stage classification is one of the primary steps for identifying sleep-related anomalies. Sleep stages are classified according to the frequency and nature of signals received during a polysomnography test. Since the early days, this has been performed manually with the help of trained technicians. However, manual scoring is often prone to error and subjectivity and requires tremendous time and effort. It is, therefore, essential to automatize this process. Several challenges from the correct selection of features remain to be faced in the machine learning-based sleep stage classification system. As an alternative, Deep Learning, capable of automatic feature extraction, proves far more reliable for this task. This experimental study analyses both techniques to compare and decide on a better approach. Three popular Machine Learning classifiers, namely Random Forest (RF), K-Nearest Neighbours (KNN), and Support Vector Machines (SVM), and a neural network comprising CNN and LSTM, have been trained on a vast base of diverse data. The proposed model reported an accuracy of 87.4% with CNN + LSTM.
In landmark localization, due to ambiguities in defining their exact position, landmark annotations may suffer from large observer variabilities, which result in uncertain annotations. To model the annotation ambiguit...
详细信息
Data-driven approaches are revolutionizing healthcare management and analysis, offering insights into patient care, operational efficiency, and future trends. The sum of information Big data’ is, can do miracles. In ...
详细信息
ISBN:
(数字)9798350366846
ISBN:
(纸本)9798350366853
Data-driven approaches are revolutionizing healthcare management and analysis, offering insights into patient care, operational efficiency, and future trends. The sum of information Big data’ is, can do miracles. In the last 20 years, it has become a topic of interest because the potential lies within is enormous. Big Data is produced, stored and examined by several industries of public and private domain with the main aim to enhance their services. Big Data comes from variety of resources in medical industry such as medical patient records, test reports, and IOT medical devices. The large amount of data is generated by medical research which is vital to public healthcare. The data must be appropriately organised and analysed to obtain useful information. Else, finding solution through big data analysis will be like looking for needle in haystack. Set of difficulties frequently occur at every phase of managing Big Data which can only be overcome by installing advanced computer systems for its analysis. This paper examines the relationship between data science and healthcare with an intent to use complex metrics to make intelligent and accurate decisions. Important areas are data-driven approaches for enhancing medical administration, recognizing new developments influencing healthcare evolution, and the methods for evaluating large sets of medical data. By incorporating the data-driven techniques, the medical industry might able to improve the patient health, optimize operations and adjust to current as well as to continuously evolving situations.
暂无评论