Aerial vehicle detection is a critical function in surveillance, disaster response, and many defense systems. In this paper, a customized model, which extends the YOLOv8 model to detect aerial vehicles with exact prec...
详细信息
In fields like human-computer interaction, healthcare, and marketing, facial and emotional recognition utilizing machine learning (ML) has found many uses. the primary method used in this study is Convolutional Neural...
详细信息
the article proposes a step-by-step solution to the problem of modeling investment policy. With a fairly extensive list of methods currently available, the authors preferred a methodology based on the digital economy ...
详细信息
ISBN:
(纸本)9798350369397;9798350369380
the article proposes a step-by-step solution to the problem of modeling investment policy. With a fairly extensive list of methods currently available, the authors preferred a methodology based on the digital economy paradigm. the digital economy is positioned as an economy based on the use of information technologies for optimal resource management, in this case, enterprises. the modeling task is solved in three stages: the formation of an algorithm, the development of a flowchart, and the writing of a computer program. the demand for modeling is dictated by the presence of a large information array, and identifying the optimal solution to the problem is impossible without optimization models. A fairly extensive sequence of calculation procedures can only be performed using computer programs.
the advancement of gesture recognition technology has been greatly influenced by the introduction of innovative devices. these devices have significantly improved hand pose recognition through more accurate tracking o...
详细信息
the objective of this research is to enhance the security of file encryption by utilizing Novel Noise Images as keys, thereby mitigating potential security threats. the proposed method employs advanced encryption stan...
详细信息
High reliance on autonomous systems necessitates efficient and reliable data exchange through Vehicle-to-Infrastructure (V2I) communication to have appropriate and stable network performance in dynamic vehicular envir...
详细信息
Nowadays, most of the transactions are online due to the high involvement of online sales and purchase purposes. Although these transactions provide convenience to their end user, meanwhile these are also the main cau...
详细信息
this paper explores and envisions a collaboration between Facial Expression Recognition (FER) and Generative Artificial Intelligence (GenAI) to create an advanced human-computer interaction (HCI) experience. Deep lear...
详细信息
Knee osteoarthritis (KOA) is a degenerative knee disease that affects the independence, quality of life, and motivation of millions of people. Early detection and accurate prediction of KOA can facilitate timely inter...
详细信息
Automatic facial ethnicity recognition from a facial image is an essential and complex challenge in face image analysis. Withthe introduction of deep convolutional neural networks, existing methods of ethnicity class...
详细信息
ISBN:
(纸本)9798350349795;9798350349788
Automatic facial ethnicity recognition from a facial image is an essential and complex challenge in face image analysis. Withthe introduction of deep convolutional neural networks, existing methods of ethnicity classification use either fully connected layers or global average pooling (GAP) layer for classification. GAP layer is employed in place of fully connected layers to minimize parameters number and computational cost. But the problem associated with GAP layer is the loss of spatial resolution. To tackle this problem, we present a new approach modified wise-srNet for ethnicity classification. Extensive experiments are conducted on three popular datasets: BUPT-transferface, FairFace and UTK using several deep convolutional neural networks with proposed architecture for performance analysis. the results from the different experiments show that our proposed method is superior to previous approaches in terms of both robustness and accuracy. Moreover, cross-dataset evaluation is also performed to investigate generalization performance and obtained 86.50% accuracy on UTK dataset demonstrating highest accuracy compared to prior methods.
暂无评论