Traditional cervical cancer diagnosis mainly relies on human papillomavirus (HPV) concentration testing. Considering that HPV concentrations vary from individual to individual and fluctuate over time, this method requires multiple tests, leading to high costs. Recently, some scholars have focused on the method of cervical cytology for diagnosis. However, cervical cancer cells have complex textural characteristics and small differences between different cell subtypes, which brings great challenges for high-precision screening of cervical cancer. In this paper, we propose a high-precision cervical cancer precancerous lesion screening classification method based on ConvNeXt, utilizing self-supervised data augmentation and ensemble learning strategies to achieve cervical cancer cell feature extraction and inter-class discrimination, respectively. We used the Deep Cervical Cytological Levels (DCCL) dataset, which includes 1167 cervical cytology specimens from participants aged 32 to 67, for algorithm training and validation. We tested our method on the DCCL dataset, and the final classification accuracy was 8.85% higher than that of previous advanced models, which means that our method has significant advantages compared to other advanced methods.
Industrial edge computing devices typically have limited computational performance. Current high-precision grasping pose detection models are often over-parameterized, resulting in computational inefficiency and making it difficult to be deployed on industrial robots. Existing efficient network design methods rely on constructing, training, and validating a large number of new networks to obtain the final target network, which also require significant computing resources. To address this issue, the Network Convergence Indicator (NCI) is proposed to realize the prediction of network accuracy before training. NCI can evaluate convergence based on the evasion of flat regions and saddle points theory. It is used to control the width and depth of the network model and optimize its innner layers. Based on these principles, NCIGrasp is proposed, which is a high-precision lightweight grasping network achieving outstanding performance on the Cornell and Jacquard datasets with only 10% of the parameters of previous advanced models. NCIGrasp has a small size of 0.69 MB and fast inference speed on server (~8ms), making it suitable for edge computing. Experimental results show that NCIGrasp has real-time grasping pose detection ability (~21ms) for unknown objects on an ordinary laptop with limited computation capability.
RFID-based pose perception can enable industrial automation applications such as industrial robot grasping. In this paper, a RFID pose estimation method based on classification algorithm and phase-position transformation model for moving objects is proposed, which converts the traditional pose estimation problem into a machine learning classification problem by dividing the direction angle value domain of the object into several classes. The phase information of multiple RFID tags attached to the object is transformed into position information using an unwrapped phase-position model, on which the input features of the classifier is constructed. A classifier based on the LightGBM framework is constructed and trained to realize the mapping between RFID phase information and the object's pose. Extensive experiments demonstrate that the proposed method in this paper can accurately estimate the pose of moving objects in real time and successfully complete the robot grasping task for objects on the conveyor belt.
Bearings are key components in mechanical equipment and remaining useful life (RUL) prediction of bearings is of great significance for stable production in industry. To predict RUL accurately, the time-frequency diagram and the long-term dependence of vibration signals have been widely researched. However, the linear frequency scale of time-frequency diagram is unfavorable to balance characteristic defect of low-frequency and natural vibration of high-frequency. Also, abrupt change vibration cause serious interference to long-term dependence. Therefore, based on theoretical analysis of bearings vibration, we introduce Mel-scale frequency cepstral coefficients (MFCC) three-channels image and propose reverse attention graph convolution network with residual convolution transformer (RCT-RAGCN). MFCC three-channels image optimizes the frequency region by logarithmic Mel scale. Reverse attention improves aggregation method of long-term dependence and alleviates the interference of abrupt change, approaches long-term dependency problems from signal difference perspective. In constructed graphs, reverse attention matrices are as edges, which is profit for aggregation of abrupt change by graph convolution network (GCN). In proposed encoder, residual convolution extract features from MFCC three-channels image and Transformer learn raw signals to prevent information loss caused in constructing MFCC. On the IEEE PHM 2012 dataset, results indicate that the proposed method outperforms five advanced models in term of accuracy. The ablation studies verify the significant role of proposed method.
In recent years, the exponential growth of online social networks as complex networks has presented challenges in expanding networks and forging new connections. Link prediction emerges as a crucial technique to anticipate future relationships among users, leveraging the current network state to address this challenge effectively. While link prediction models on monoplex networks have a well-established history, the exploration of similar tasks on multilayer networks has garnered considerable attention. Extracting topological and multimodal features for weighting links can improve link prediction in weighted complex networks. Meanwhile, establishing reliable and trustworthy paths between users is a useful way to create metrics that convert unweighted to weighted similarity. The local random walk is a widely used technique for predicting links in weighted monoplex networks. The aim of this paper is to develop a semi-local random walk over reliable paths to improve link prediction on a multilayer social network as a complex network, which is denoted as Reliable Multiplex semi-Local Random Walk (RMLRW). RMLRW leverages the semi-local random walk technique over reliable paths, integrating intra-layer and inter-layer information from multiplex features to conduct a trustworthy biased random walk for predicting new links within a target layer of multilayer networks. In order to make RMLRW scalable, we develop a semi-local random walk-based network embedding to represent the network in a lower-dimensional space while preserving its original characteristics. Extensive experimental studies on several real-world multilayer networks demonstrate the performance assurance of RMLRW compared to equivalent methods. Specifically, RMLRW improves the average f-measure of the link prediction by 3.2% and 2.5% compared to SEM-Path and MLRW, respectively.