Multiclass segmentation of hand bones in X-ray images is crucial for various medical applications, including diagnostic assistance and surgical planning. This study presents BoneSegNet, a deep learning-based segmentat...
详细信息
ISBN:
(数字)9798350364569
ISBN:
(纸本)9798350364576
Multiclass segmentation of hand bones in X-ray images is crucial for various medical applications, including diagnostic assistance and surgical planning. This study presents BoneSegNet, a deep learning-based segmentation model designed to classify hand bones into seven distinct anatomical classes: Distal phalanges, Intermediate phalanges, Proximal phalanges, Metacarpals, Carpals, Ulna, and Radius. Our model, BoneSegNet leverages the strengths of both convolutional neural networks (CNNs) and transformer encoders incorporating the UNet architecture. Our model architecture begins with patch and positional embeddings of the input image, followed by a series of transformer encoder layers to capture long-range dependencies and contextual information. The transformer encoder is equipped with multi-head attention mechanisms and multi-layer perceptron’s (MLPs) with GELU activation and dropout for regularization. Skip connections at specific layers are stored and later utilized in the decoder. The decoding path consists of a series of deconvolution and convolution blocks designed to progressively upscale the feature maps while incorporating skip connections from the encoder. This hierarchical structure ensures detailed spatial information is preserved and enhanced. The final output layer employs a 1x1 convolution with a sigmoid activation function to produce the multiclass segmentation map. The model is trained and evaluated on a dataset of hand X-ray images. It achieves an average accuracy of 99.25%, an average Dice score of 0.81, an average mIoU score of 0.70, sensitivity of 0.79, and specificity of 0.997. These results demonstrate the model’s effectiveness in accurately segmenting and classifying hand bones, making it a valuable tool for medical professionals.
Hearing impairment is a widespread health condition that affects millions of people worldwide. It refers to the lack of or low functionality of the hearing organs, resulting in difficulties in effective communication ...
Hearing impairment is a widespread health condition that affects millions of people worldwide. It refers to the lack of or low functionality of the hearing organs, resulting in difficulties in effective communication and participation in everyday conversations. This limitation can negatively impact various aspects of the social and personal lives of the hearing-impaired individuals. Moreover, the high cost and limited availability of assistive devices further restrict accessibility for those with limited financial resources. To address these challenges, this study develops an affordable wearable device equipped with speech and voice recognition technologies to enhance the accessibility of spoken language for individuals with hearing impairment. OpenAI Whisper is employed for real-time conversion of spoken words into written text. Voice recognition, utilizing Mel Frequency Cepstral Coefficients (MFCCs) and Random Forest (RF) classifier, identifies and distinguishes speakers within the user's family and friends zone. The recognized spoken words from the speech recognition process and the speaker's identity from the voice recognition process are combined and displayed on a screen to the hearing impaired. By combining speech and voice recognition technologies, the proposed device breaks communication barriers for individuals with hearing impairment and enhance their overall communication experience.
As electric vehicle adoption increases, charging stations will face greater demand and have greater energy needs. This paper outlines a system for reserving electric vehicle charging, using a reservation called an ene...
As electric vehicle adoption increases, charging stations will face greater demand and have greater energy needs. This paper outlines a system for reserving electric vehicle charging, using a reservation called an energy packet. The proposed energy packets are composed of the charging duration and the charging energy. Using energy packet reservations from EV users as part of the inputs, this paper developed a charging management system through a charging optimization algorithm to achieve reduced total charging costs for each electric vehicle. A case study is presented to demonstrate the effectiveness of the proposed method in reducing charging costs, peak demand, and greenhouse gas emissions of a charging station.
The Intelligent Transportation System (ITS) has become a part of smart cities and road safety. Based on communication systems technologies, ITS can solve several road issues, such as accidents. A vehicular ad-hoc netw...
详细信息
We present our work on leveraging low-frame-rate monochrome (blue light) videos of fingertips, captured with an off-the-shelf fingerprint capture device, to extract vital signs and identify users. These videos utilize...
详细信息
The efficient operation of the unified Integrated Sensing and Communication (ISAC) – Mobile Edge Computing (MEC) systems is important for enhancing data sensing, communication, and computation processes in next-gener...
详细信息
ISBN:
(数字)9798350351255
ISBN:
(纸本)9798350351262
The efficient operation of the unified Integrated Sensing and Communication (ISAC) – Mobile Edge Computing (MEC) systems is important for enhancing data sensing, communication, and computation processes in next-generation wireless systems. Despite prior research focusing on these systems, little attention has been given to optimizing the device-edge server associations. This paper addresses this gap by introducing the novel two-stage device-edge server association Synergia framework. Firstly, representative utility functions capture the characteristics of the devices and MEC servers by jointly considering their sensing, communication, and computation characteristics. Secondly, the Estimated Synergia framework leverages the Matching Theory to rapidly determine an initial device-server matching by disregarding the devices’ externalities, i.e., the matching decisions of other devices. Thirdly, the Accurate Synergia model refines and improves this matching by using the coalition formation games, while considering the devices’ externalities in optimizing the utilities of both the devices and the MEC servers. Extensive numerical evaluations demonstrate the Synergia’s operational efficiency and scalability, outperforming reinforcement learningbased approaches. Also, a real-world application involving car accident detection validates its applicability.
In the search for more effective education, there is a widespread effort to develop better approaches to personalize student education. Unassisted, educators often do not have time or resources to personally support e...
详细信息
Reinforcement learning yields a feedback controller that achieves specific control goal (which is often translated as a reward function). However, it often suffers from the Sim2Real gap, and domain randomization is kn...
详细信息
Multiple input multiple output(MIMO) radar systems have seen considerable use due to their ability to simulate a large antenna array with few physical antenna elements. The performance of microwave sensors can benefit...
详细信息
ISBN:
(数字)9798350349375
ISBN:
(纸本)9798350349382
Multiple input multiple output(MIMO) radar systems have seen considerable use due to their ability to simulate a large antenna array with few physical antenna elements. The performance of microwave sensors can benefit greatly from MIMO techniques. This paper presents a simulation study to determine the feasibility of MIMO radars leveraging the stepped-frequency continuous wave(SFCW) transmit scheme and investigates the possibility of utilizing Doppler division multiple access(DDMA) for MIMO SFCW radar. Although DDMA is a well-known slow-time phase coding method to synthesize frequency division and generate orthogonal MIMO waveforms, it has not been leveraged in the context of SFCW radar. Therefore, this paper compares SFCW with the more common frequency modulated continuous wave(FMCW) technique to compare the merits and demerits of each transmit scheme. A MATLAB simulation is presented to analyze the possibility of DDMA SFCW radars.
Network data plays a crucial role in various real-world applications, as connections between entities can be represented and analyzed through graphs. These include various types, such as social, information, and techn...
详细信息
ISBN:
(数字)9798331531904
ISBN:
(纸本)9798331531911
Network data plays a crucial role in various real-world applications, as connections between entities can be represented and analyzed through graphs. These include various types, such as social, information, and technical networks. However, the complex topologies of these networks present challenges in converting graph data into machine-readable vector formats. Existing models, such as Graph Neural Networks, Graph Attention Networks, and node2vec, have made strides in graph embeddings. For edge-related tasks, models such as node2vec typically use indirect methods like concatenating node vectors to represent edges. This approach is useful for tasks like link prediction. In this paper, we present LineDi2vec, a novel approach that improves the Node2vec embedding method by utilizing a line graph. The proposed LineDi2vec not only generalizes the original graphs, transforming the relationships between edges and nodes but also maintains the original graphs' topological integrity for effective node embedding by Node2vec. We evaluated LineDi2vec on four real-world datasets, focusing on link prediction. The results demonstrate that LineDi2vec outperforms traditional node concatenation methods.
暂无评论