Brain-computer interfaces (BCIs) enable direct communication between the brain and external devices. This review highlights the core decoding algorithms that enable multimodal BCIs, including a dissection of the eleme...
详细信息
Interactive medical image segmentation methods have become increasingly popular in recent years. These methods combine manual labeling and automatic segmentation, reducing the workload of annotation while maintaining ...
详细信息
Most proposed memristor-based circuits of associative memory consider various mechanisms in only one associative memory. Few works on circuit design of sequential associative memory have been reported. In this paper, ...
详细信息
Accurate and automated gland segmentation on pathological images can assist pathologists in diagnosing the malignancy of colorectal adenocarcinoma. However, due to various gland shapes, severe deformation of malignant...
详细信息
The brain is the most sophisticated and complex organ in the human body. Nowadays, diagnosing complex and diverse brain diseases is a hot topic. Alzheimer's Disease (AD), Autism Spectrum Disorder (ASD), and others...
The brain is the most sophisticated and complex organ in the human body. Nowadays, diagnosing complex and diverse brain diseases is a hot topic. Alzheimer's Disease (AD), Autism Spectrum Disorder (ASD), and others are common brain diseases. With the increased usage of Artificial Intelligence (AI) in medical image analysis, the endeavor to make AI comprehend brain images for assisting doctors in making objective diagnoses of brain diseases has gained considerable attention. Resting-state functional magnetic resonance imaging (rs-fMRI) is a widely used tool for diagnosing and analyzing brain ***, the Pearson correlation coefficient (PCC) method constructs a dynamic functional connectivity (dFC) network using a fixed window size. However, this method has limitations as it is challenging to extract potential high-level features from the dFC and determine local feature relevance. This paper presents a method for constructing dFC based on multi-scale sliding windows and proposes Multi-scale Convolutional Neural Networks (MsCNN) for learning and analyzing dFC at various scales. Finally, a deep fusion of features learned at different scales is employed for the diagnostic classification of brain *** to the rs-fMRI dataset from ADNI, the classification accuracy was 84.0% for eMCI/lMCI, 84.4% for lMCI/AD, and 58.4% for NC/eMCI/lMCI/AD. Applied to rs-fMRI data from ABIDE, the accuracy was 74.7% for ASD/NC. The proposed method exhibits robust classification performance and introduces a new approach to diagnosing other brain diseases.
With the rapid development of smart agriculture, the demand for pig monitoring has been increasing, particularly in disease prevention and behavioral analysis. Real-time monitoring of pig behavior is essential for imp...
详细信息
ISBN:
(数字)9798350389807
ISBN:
(纸本)9798350389814
With the rapid development of smart agriculture, the demand for pig monitoring has been increasing, particularly in disease prevention and behavioral analysis. Real-time monitoring of pig behavior is essential for improving farming efficiency and ensuring animal welfare. However, challenges such as pigs’ varying body sizes, movement, and complex environments make accurate detection and tracking difficult. To address these issues, this paper proposes an improved YOLOv5 model, named PLM-YOLOv5, specifically designed for pig detection in livestock monitoring. First, we add an additional detection head to enhance the model’s capability to detect smaller pigs effectively. Next, we integrate the Convolutional Block Attention Module (CBAM) to enhance feature extraction, particularly in handling complex backgrounds. Finally, we utilize a self-collected dataset of over 1,400 images for model training and validation, ensuring the model adapts to real-world livestock environments. Experimental results show that PLM-YOLOv5 achieves significant improvements in pig detection tasks, with the mean Average Precision (mAP) increasing by approximately 9% compared to the original YOLOv5. Additionally, training and validation loss curves demonstrate the model's stability and convergence. These results confirm the effectiveness and reliability of PLM-YOLOv5 for real-world applications, offering a robust solution for smart agriculture.
Accurate and automated gland segmentation on pathological images can assist pathologists in diagnosing the malignancy of colorectal adenocarcinoma. However, due to various gland shapes, severe deformation of malignant...
Accurate and automated gland segmentation on pathological images can assist pathologists in diagnosing the malignancy of colorectal adenocarcinoma. However, due to various gland shapes, severe deformation of malignant glands, and overlapping adhesions between glands. Gland segmentation has always been very challenging. To address these problems, we propose a DEA model. This model consists of two branches: the backbone encoding and decoding network and the local semantic extraction network. The backbone encoding and decoding network extracts advanced Semantic features, uses the proposed feature decoder to restore feature space information, and then enhances the boundary features of the gland through boundary enhancement attention. The local semantic extraction network uses the pre-trained DeepLabv3+ as a Local semantic-guided encoder to realize the extraction of edge features. Experimental results on two public datasets, GlaS and CRAG, confirm that the performance of our method is better than other gland segmentation methods.
Background and ObjectiveThe instance segmentation of impacted tooth in the oral panoramic X-ray images is research hot. However, impacted tooth in panoramic X-Ray images lead to teeth deformities, low contrast between...
详细信息
There are some challenges in multimodal medical image segmentation. Based on this, the Model-Data Co-driven U-Net Segmentation Network for Multimodal Lung Tumor images is proposed. About "How to extract edge feat...
详细信息
Generalized Zero-Shot Learning (GZSL) aims to leverage a classifier trained on seen classes to categorize instances from both seen and unseen classes. Several approaches have been introduced to synthesize visual featu...
Generalized Zero-Shot Learning (GZSL) aims to leverage a classifier trained on seen classes to categorize instances from both seen and unseen classes. Several approaches have been introduced to synthesize visual features that simulate those of unseen classes for training classifiers. However, existing methods only emphasize the distributional relationships between synthesized and real features, while neglecting the inter-class relationships among the synthesized features. Consequently, synthesized visual features exhibit significant loose intra-class distributions and numerous outliers. Furthermore, the generator trained solely on seen classes tend to overfit these classes. In this paper, a Semantic Fusion and Contrastive Generation (SFCG) framework is proposed for GZSL. Specifically, a visual-semantic contrastive generation method and a visual features similarity loss are explored to address the challenges of loose intra-class distribution and outliers in synthesized visual features. Moreover, semantic attributes are fused to create novel and diverse semantic instances for training a balanced generator. The SFCG model is evaluated on four widely-used ZSL benchmark datasets: CUB, FLO, AWA2, and SUN. It achieves harmonic mean accuracies of 68.4% on CUB, 71.3% on FLO, 73.4% on AWA2, and 45.2% on SUN, demonstrating the efficacy of the proposed method.
暂无评论