Search Results (5)

Post-operative early recurrence (ER) of hepatocellular carcinoma (HCC) is a major cause of mortality. Predicting ER before treatment can guide treatment and follow-up protocols. Deep learning frameworks, known for their superior performance, are widely used in medical imaging. However, they face challenges due to limited annotated data. We propose a multi-task pre-training method using self-supervised learning with medical images for predicting the ER of HCC. This method involves two pretext tasks: phase shuffle, focusing on intra-image feature representation, and case discrimination, focusing on inter-image feature representation. The effectiveness and generalization of the proposed method are validated through two different experiments. In addition to predicting early recurrence, we also apply the proposed method to the classification of focal liver lesions. Both experiments show that the multi-task pre-training model outperforms existing pre-training (transfer learning) methods with natural images, single-task self-supervised pre-training, and DINOv2. Full article

Forest fires pose significant hazards to ecological environments and economic society. The detection of forest fire smoke can provide crucial information for the suppression of early fires. Previous detection models based on deep learning have been limited in detecting small smoke and smoke with smoke-like interference. In this paper, we propose a lightweight model for forest fire smoke detection that is suitable for UAVs. Firstly, a smoke dataset is created from a combination of forest smoke photos obtained through web crawling and enhanced photos generated by using the method of synthesizing smoke. Secondly, the GSELAN and GSSPPFCSPC modules are built based on Ghost Shuffle Convolution (GSConv), which efficiently reduces the number of parameters in the model and accelerates its convergence speed. Next, to address the problem of indistinguishable feature boundaries between clouds and smoke, we integrate coordinate attention (CA) into the YOLO feature extraction network to strengthen the extraction of smoke features and attenuate the background information. Additionally, we use Content-Aware Reassembly of FEatures (CARAFE) upsampling to expand the receptive field in the feature fusion network and fully exploit the semantic information. Finally, we adopt SCYLLA-Intersection over Union (SIoU) loss as a replacement for the original loss function in the prediction phase. This substitution leads to improved convergence efficiency and faster convergence. The experimental results demonstrate that the LMDFS model proposed for smoke detection achieves an accuracy of 80.2% with a 5.9% improvement compared to the baseline and a high number of Frames Per Second (FPS)—63.4. The model also reduces the parameter count by 14% and Giga FLoating-point Operations Per second (GFLOPs) by 6%. These results suggest that the proposed model can achieve a high accuracy while requiring fewer computational resources, making it a promising approach for practical deployment in applications for detecting smoke. Full article

Coverage is a fundamental issue in wireless sensor networks (WSNs). It plays a important role in network efficiency and performance. When sensor nodes are randomly scattered in the network environment, an ON/OFF scheduling mechanism can be designed for these nodes to ensure network coverage and increase the network lifetime. In this paper, we propose an appropriate and optimal area coverage method. The proposed area coverage scheme includes four phases: (1) Calculating the overlap between the sensing ranges of sensor nodes in the network. In this phase, we present a novel, distributed, and efficient method based on the digital matrix so that each sensor node can estimate the overlap between its sensing range and other neighboring nodes. (2) Designing a fuzzy scheduling mechanism. In this phase, an ON/OFF scheduling mechanism is designed using fuzzy logic. In this fuzzy system, if a sensor node has a high energy level, a low distance to the base station, and a low overlap between its sensing range and other neighboring nodes, then this node will be in the ON state for more time. (3) Predicting the node replacement time. In this phase, we seek to provide a suitable method to estimate the death time of sensor nodes and prevent possible holes in the network, and thus the data transmission process is not disturbed. (4) Reconstructing and covering the holes created in the network. In this phase, the goal is to find the best replacement strategy of mobile nodes to maximize the coverage rate and minimize the number of mobile sensor nodes used for covering the hole. For this purpose, we apply the shuffled frog-leaping algorithm (SFLA) and propose an appropriate multi-objective fitness function. To evaluate the performance of the proposed scheme, we simulate it using NS2 simulator and compare our scheme with three methods, including CCM-RL, CCA, and PCLA. The simulation results show that our proposed scheme outperformed the other methods in terms of the average number of active sensor nodes, coverage rate, energy consumption, and network lifetime. Full article

The differentiation of autoimmune pancreatitis (AIP) and pancreatic ductal adenocarcinoma (PDAC) poses a relevant diagnostic challenge and can lead to misdiagnosis and consequently poor patient outcome. Recent studies have shown that radiomics-based models can achieve high sensitivity and specificity in predicting both entities. However, radiomic features can only capture low level representations of the input image. In contrast, convolutional neural networks (CNNs) can learn and extract more complex representations which have been used for image classification to great success. In our retrospective observational study, we performed a deep learning-based feature extraction using CT-scans of both entities and compared the predictive value against traditional radiomic features. In total, 86 patients, 44 with AIP and 42 with PDACs, were analyzed. Whole pancreas segmentation was automatically performed on CT-scans during the portal venous phase. The segmentation masks were manually checked and corrected if necessary. In total, 1411 radiomic features were extracted using PyRadiomics and 256 features (deep features) were extracted using an intermediate layer of a convolutional neural network (CNN). After feature selection and normalization, an extremely randomized trees algorithm was trained and tested using a two-fold shuffle-split cross-validation with a test sample of 20% (n = 18) to discriminate between AIP or PDAC. Feature maps were plotted and visual difference was noted. The machine learning (ML) model achieved a sensitivity, specificity, and ROC-AUC of 0.89 ± 0.11, 0.83 ± 0.06, and 0.90 ± 0.02 for the deep features and 0.72 ± 0.11, 0.78 ± 0.06, and 0.80 ± 0.01 for the radiomic features. Visualization of feature maps indicated different activation patterns for AIP and PDAC. We successfully trained a machine learning model using deep feature extraction from CT-images to differentiate between AIP and PDAC. In comparison to traditional radiomic features, deep features achieved a higher sensitivity, specificity, and ROC-AUC. Visualization of deep features could further improve the diagnostic accuracy of non-invasive differentiation of AIP and PDAC. Full article

The prediction aptitude of an artificial neural network (ANN) is improved by incorporating two novel metaheuristic techniques, namely, the shuffled frog leaping algorithm (SFLA) and wind-driven optimization (WDO), for the purpose of soil shear strength (simply called shear strength) simulation. Soil information of the Trung Luong national expressway project (Vietnam) including depth of the sample (m), percentage of sand, percentage of silt, percentage of clay, percentage of moisture content, wet density (kg/m³), liquid limit (%), plastic limit (%), plastic index (%), liquidity index, and the shear strength (kPa) was collocated through a field survey. After constructing the hybrid ensembles of SFLA–ANN and WDO–ANN, both models were optimized in terms of complexity using a population-based trial-and error-scheme. The learning quality of the ANN was compared with both improved versions to examine the effect of the used metaheuristic techniques. In this phase, the training error dropped by 14.25% and 28.25% by applying the SFLA and WDO, respectively. This reflects a significant improvement in pattern recognition ability of the ANN. The results of the testing data revealed 25.57% and 39.25% decreases in generalization (i.e., testing) error. Moreover, the correlation between the measured and predicted shear strengths (i.e., the coefficient of determination) rose from 0.82 to 0.89 and 0.92, which indicates the efficiency of both SFLA and WDO metaheuristic techniques in optimizing the ANN. Full article