Enhanced multi view 3D reconstruction with improved MVSNet.

Although 3D reconstruction has been widely used in many fields as a key component of environment perception, existing technologies still have the potential for further improvement in 3D scene reconstruction. We propose an improved reconstruction algorithm based on the MVSNet network architecture. To glean richer pixel details from images, we suggest deploying a DE module integrated with a residual framework, which supplants the prevailing feature extraction mechanism. The DE module uses ECA-Net and dilated convolution to expand the receptive field range, performing feature splicing and fusion through the residual structure to retain the global information of the original image. Moreover, harnessing attention mechanisms refines the 3D cost volume's regularization process, bolstering the integration of information across multi-scale feature volumes, consequently enhancing depth estimation precision. When assessed our model using the DTU dataset, findings highlight the network's 3D reconstruction scoring a completeness (comp) of 0.411 mm and an overall quality of 0.418 mm. This performance is higher than that of traditional methods and other deep learning-based methods. Additionally, the visual representation of the point cloud model exhibits marked advancements. Trials on the Blended MVS dataset signify that our network exhibits commendable generalization prowess.

Screening of postoperative cerebral hyperperfusion syndrome in moyamoya disease: a three-dimensional pulsed arterial-spin labeling magnetic resonance imaging approach.

Introduction: Moyamoya disease (MMD) is associated with a risk of postoperative cerebral hyperperfusion syndrome (CHS) after revascularization surgery. This study aimed to explore the feasibility of using three-dimensional pulsed arterial spin labeling (3D PASL) and phase contrast (PC) magnetic resonance imaging (MRI) for predicting CHS occurrence in patients with MMD before revascularization surgery. Methods: Overall, 191 adult patients (207 hemispheres) with MMD who underwent combined revascularization surgery were included in this study. Preoperative 3D PASL-MRI and PC-MRI were performed before surgery. The PASL-MRI data were analyzed using SPM12. Patient clinical information, average flow, and preoperative cerebral blood flow (CBF) were compared between the non-CHS and CHS groups. Results: Among the patients, 45 (21.74%) developed CHS after revascularization surgery. No significant differences were noted in age, sex, clinical symptoms, hypertension, diabetes, surgical side, or history of revascularization surgery between the non-CHS and CHS groups. However, the average flow in the superficial temporal artery was significantly lower in the CHS group than in the non-CHS group (p < 0.05). Furthermore, 11 clusters of preoperative CBF values were significantly greater in the CHS group than in the non-CHS group [p < 0.05, false discovery rate (FDR) corrected]. A significant correlation was also observed between the preoperative time-to-flight MR angiography (MRA) scores and CBF values in patients with MMD (p < 0.05). Conclusion: Compare patients with lower preoperative CBF and higher preoperative average flow in the STA, patients with higher preoperative CBF and lower preoperative average flow in the STA are more likely to develop postoperative CHS Preoperative PASL-MRI and PC-MRI examinations may help to screen patients at high risk of developing CHS after revascularization surgery.

EFC-YOLO: An Efficient Surface-Defect-Detection Algorithm for Steel Strips.

The pursuit of higher recognition accuracy and speed with smaller model sizes has been a major research topic in the detection of surface defects in steel. In this paper, we propose an improved high-speed and high-precision Efficient Fusion Coordination network (EFC-YOLO) without increasing the model's size. Since modifications to enhance feature extraction in shallow networks tend to affect the speed of model inference, in order to simultaneously ensure the accuracy and speed of detection, we add the improved Fusion-Faster module to the backbone network of YOLOv7. Partial Convolution (PConv) serves as the basic operator of the module, which strengthens the feature-extraction ability of shallow networks while maintaining speed. Additionally, we incorporate the Shortcut Coordinate Attention (SCA) mechanism to better capture the location information dependency, considering both lightweight design and accuracy. The de-weighted Bi-directional Feature Pyramid Network (BiFPN) structure used in the neck part of the network improves the original Path Aggregation Network (PANet)-like structure by adding step branches and reducing computations, achieving better feature fusion. In the experiments conducted on the NEU-DET dataset, the final model achieved an 85.9% mAP and decreased the GFLOPs by 60%, effectively balancing the model's size with the accuracy and speed of detection.

KRP-DS: A Knowledge Graph-Based Dialogue System with Inference-Aided Prediction.

With the popularity of ChatGPT, there has been increasing attention towards dialogue systems. Researchers are dedicated to designing a knowledgeable model that can engage in conversations like humans. Traditional seq2seq dialogue models often suffer from limited performance and the issue of generating safe responses. In recent years, large-scale pretrained language models have demonstrated their powerful capabilities across various domains. Many studies have leveraged these pretrained models for dialogue tasks to address concerns such as safe response generation. Pretrained models can enhance responses by carrying certain knowledge information after being pre-trained on large-scale data. However, when specific knowledge is required in a particular domain, the model may still generate bland or inappropriate responses, and the interpretability of such models is poor. Therefore, in this paper, we propose the KRP-DS model. We design a knowledge module that incorporates a knowledge graph as external knowledge in the dialogue system. The module utilizes contextual information for path reasoning and guides knowledge prediction. Finally, the predicted knowledge is used to enhance response generation. Experimental results show that our proposed model can effectively improve the quality and diversity of responses while having better interpretability, and outperforms baseline models in both automatic and human evaluations.

A lightweight model for efficient identification of plant diseases and pests based on deep learning.

Plant diseases and pests have always been major contributors to losses that occur in agriculture. Currently, the use of deep learning-based convolutional neural network models allows for the accurate identification of different types of plant diseases and pests. To enable more efficient identification of plant diseases and pests, we design a novel network architecture called Dise-Efficient based on the EfficientNetV2 model. Our experiments demonstrate that training this model using a dynamic learning rate decay strategy can improve the accuracy of plant disease and pest identification. Furthermore, to improve the model's generalization ability, transfer learning is incorporated into the training process. Experimental results indicate that the Dise-Efficient model boasts a compact size of 13.3 MB. After being trained using the dynamic learning rate decay strategy, the model achieves an accuracy of 99.80% on the Plant Village plant disease and pest dataset. Moreover, through transfer learning on the IP102 dataset, which represents real-world environmental conditions, the Dise-Efficient model achieves a recognition accuracy of 64.40% for plant disease and pest identification. In light of these results, the proposed Dise-Efficient model holds great potential as a valuable reference for the deployment of automatic plant disease and pest identification applications on mobile and embedded devices in the future.

Decision Method of Optimal Needle Insertion Angle for Dorsal Hand Intravenous Robot.

In the context of COVID-19, the research on various aspects of the venipuncture robot field has become increasingly hot, but there has been little research on robotic needle insertion angles, primarily performed at a rough angle. This will increase the rate of puncture failure. Furthermore, there is sometimes significant pain due to the patients' differences. This paper investigates the optimal needle entry angle decision for a dorsal hand intravenous injection robot. The dorsal plane of the hand was obtained by a linear structured light scan, which was used as a basis for calculating the needle entry angle. Simulation experiments were also designed to determine the optimal needle entry angle. Firstly, the linear structured optical system was calibrated and optimized, and the error function was constructed and solved iteratively by the optimization method to eliminate measurement error. Besides, the dorsal hand was scanned to obtain the spatial point clouds of the needle entry area, and the least squares method was used to fit it to obtain the dorsal hand plane. Then, the needle entry angle was calculated based on the needle entry area plane. Finally, the changes in the penetration force under different needle entry angles were analyzed to determine the optimal needle insertion angle. According to the experimental results, the average error of the optimized structured light plane position was about 0.1 mm, which meets the needs of the project, and a large angle should be properly selected for needle insertion during the intravenous injection.

Determination of the location of the needle entry point based on an improved pruning algorithm.

Since the emergence of new coronaviruses and their variant virus, a large number of medical resources around the world have been put into treatment. In this case, the purpose of this article is to develop a handback intravenous intelligence injection robot, which reduces the direct contact between medical staff and patients and reduces the risk of infection. The core technology of hand back intravenous intelligent robot is a handlet venous vessel detection and segmentation and the position of the needle point position decision. In this paper, an image processing algorithm based on U-Net improvement mechanism (AT-U-Net) is proposed for core technology. It is investigated using a self-built dorsal hand vein database and the results show that it performs well, with an F1-score of 93.91%. After the detection of a dorsal hand vein, this paper proposes a location decision method for the needle entry point based on an improved pruning algorithm (PT-Pruning). The extraction of the trunk line of the dorsal hand vein is realized through this algorithm. Considering the vascular cross-sectional area and bending of each vein injection point area, the optimal injection point of the dorsal hand vein is obtained via a comprehensive decision-making process. Using the self-built dorsal hand vein injection point database, the accuracy of the detection of the effective injection area reaches 96.73%. The accuracy for the detection of the injection area at the optimal needle entry point is 96.50%, which lays a foundation for subsequent mechanical automatic injection.

Improving extractive document summarization with sentence centrality.

Extractive document summarization (EDS) is usually seen as a sequence labeling task, which extracts sentences from a document one by one to form a summary. However, extracting sentences separately ignores the relationship between the sentences and documents. One solution is to use sentence position information to enhance sentence representation, but this will cause the sentence-leading bias problem, especially in news datasets. In this paper, we propose a novel sentence centrality for the EDS task to address these two problems. The sentence centrality is based on directed graphs, while reflecting the sentence-document relationship, it also reflects the sentence position information in the document. We implicitly strengthen the relevance of sentences and documents by using sentence centrality to enhance sentence representation. Notably, we replaced the sentence position information with sentence centrality to reduce sentence-leading bias without causing model performance degradation. Experiments on the CNN/Daily Mail dataset showed that EDS models with sentence centrality significantly improved compared with baseline models.

Graph convolutional networks with hierarchical multi-head attention for aspect-level sentiment classification.

Aspect-level sentiment classification has been widely used by researchers as a fine-grained sentiment classification task to predict the sentiment polarity of specific aspect words in a given sentence. Previous studies have shown relatively good experimental results using graph convolutional networks, so more and more approaches are beginning to exploit sentence structure information for this task. However, these methods do not link aspect word and context well. To address this problem, we propose a method that utilizes a hierarchical multi-head attention mechanism and a graph convolutional network (MHAGCN). It fully considers syntactic dependencies and combines semantic information to achieve interaction between aspect words and context. To fully validate the effectiveness of the method proposed in this paper, we conduct extensive experiments on three benchmark datasets, which, according to the experimental results, show that the method outperforms current methods.

Non-Invasive Evaluation of Cerebral Hemodynamic Changes After Surgery in Adult Patients With Moyamoya Using 2D Phase-Contrast and Intravoxel Incoherent Motion MRI.

Objective: To explore the feasibility of 2D phase-contrast MRI (PC-MRI) and intravoxel incoherent motion (IVIM) MRI to assess cerebrovascular hemodynamic changes after surgery in adult patients with moyamoya disease (MMD). Methods: In total, 33 patients with MMD who underwent 2D PC-MRI and IVIM examinations before and after surgery were enrolled. Postsurgical changes in peak and average velocities, average flow, forward volume, and the area of superficial temporal (STA), internal carotid (ICA), external carotid (ECA), and vertebral (VA) arteries were evaluated. The microvascular perfusion status was compared between the hemorrhage and non-hemorrhage groups. Results: The peak velocity, average flow, forward volume, area of both the ipsilateral STA and ECA, and average velocity of the ipsilateral STA were increased (p < 0.05). The average flow and forward volume of both the ipsilateral ICA and VA and the area of the ipsilateral VA were increased (p < 0.05). The peak velocity, average velocity, average flow and forward volume of the contralateral STA, and the area of the contralateral ICA and ECA were also increased (p < 0.05), whereas the area of the contralateral VA was decreased (p < 0.05). The rf value of the ipsilateral anterior cerebral artery (ACA) supply area was increased (p < 0.05) and more obvious in the non-hemorrhage group (p < 0.05). Conclusion: Two-dimensional PC-MRI and IVIM may have the potential to non-invasively evaluate cerebrovascular hemodynamic changes after surgery in patients with MMD. An improvement in the microvascular perfusion status is more obvious in patients with ischemic MMD than in patients with hemorrhagic MMD.

Intravoxel Incoherent Motion Magnetic Resonance Imaging Used in Preoperative Screening of High-Risk Patients With Moyamoya Disease Who May Develop Postoperative Cerebral Hyperperfusion Syndrome.

Objective: This study aimed to investigate the feasibility of preoperative intravoxel incoherent motion (IVIM) MRI for the screening of high-risk patients with moyamoya disease (MMD) who may develop postoperative cerebral hyperperfusion syndrome (CHS). Methods: This study composed of two parts. In the first part 24 MMD patients and 24 control volunteers were enrolled. IVIM-MRI was performed. The relative pseudo-diffusion coefficient, perfusion fraction, apparent diffusion coefficient, and diffusion coefficient (rD*, rf, rADC, and rD) values of the IVIM sequence were compared according to hemispheres between MMD patient and healthy control groups. In the second part, 98 adult patients (124 operated hemispheres) with MMD who underwent surgery were included. Preoperative IVIM-MRI was performed. The rD*, rf, rADC, rD, and rfD* values of the IVIM sequence were calculated and analyzed. Operated hemispheres were divided into CHS and non-CHS groups. Patients' age, sex, Matsushima type, Suzuki stage, and IVIM-MRI examination results were compared between CHS and non-CHS groups. Results: Only the rf value was significantly higher in the healthy control group than in the MMD group (P < 0.05). Out of 124 operated hemispheres, 27 were assigned to the CHS group. Patients with clinical presentation of Matsushima types I-V were more likely to develop CHS after surgery (P < 0.05). The rf values of the ipsilateral hemisphere were significantly higher in the CHS group than in the non-CHS group (P < 0.05). The rfD* values of the ACA and MCA supply areas of the ipsilateral hemisphere were significantly higher in the CHS group than in the non-CHS group (P < 0.05). Only the rf value of the anterior cerebral artery supply area in the contralateral hemisphere was higher in the CHS group than in the non-CHS group (P < 0.05). The rf values of the middle and posterior cerebral artery supply areas and the rD, rD*, and rADC values of the both hemispheres were not significantly different between the CHS and non-CHS groups (P > 0.05). Conclusion: Preoperative non-invasive IVIM-MRI analysis, particularly the f-value of the ipsilateral hemisphere, may be helpful in predicting CHS in adult patients with MMD after surgery. MMD patients with ischemic onset symptoms are more likely to develop CHS after surgery.

Risk model based on MRI fusion biopsy characteristics predicts biochemical recurrence after radical prostatectomy.

BACKGROUND: To determine the prostate cancer biochemical recurrence-related fusion biopsy characteristics before radical surgery and to establish the risk prediction model of biochemical recurrence of prostate cancer. METHODS: Three hundred and four patients undergoing radical surgery for prostate cancer at Huadong Hospital affiliated to Fudan University between 2009 and 2020 for preoperative magnetic resonance imaging (MRI) before biopsy with suspicious prostate cancer lesions. Each case was followed by a 10 + x needle combination of targeted biopsy (intentional or robotic fusion) with systematic biopsy. Prostate-specific antigen levels were measured at 1, 3, and 6 months postoperatively, followed by reexamination every 6 months. Survival analysis was performed by the Kaplan-Meier method, univariate and multivariate analysis by Cox, and Logistic risk regression models. RESULTS: Higher Prostate Imaging Reporting And Data System (PI-RADS) scores (p < 0.001), suspicious extracapsular invasion (p < 0.001), and seminal vesicle invasion (p < 0.001) on MRI, the largest lesion diameter on MRI (p = 0.006), higher biopsy International Society of Urological Pathology (ISUP) grade group (p < 0.001) related to higher biochemical recurrence rates, higher pathological staging (p < 0.001), and a greater probability of local lymph node metastasis (p < 0.001). We accurately predicted the biochemical recurrence of prostate cancer after radical surgery based on preoperative features including the long diameter of the largest MRI lesion more than 23 mm, seminal vesicle invasion on MRI, and targeted fusion biopsy ISUP grade >3 Risk stratified classification (AUC = 0.729, p < 0.001). In our cohort, this risk stratification had a larger area under the curve than predictive models based only on magnetic resonance parameters and traditional risk scores. CONCLUSIONS: In this cohort, seminal vesicle invasion on MRI, the long diameter of the largest MRI lesion, and targeted fusion biopsy ISUP grade grope are significantly predictive of pathologic features and biochemical recurrence after prostate surgery. The risk stratification integrating the three parameters could better predict the biochemical recurrence than the traditional model.