Search

Background: Nerve transfers represent a new paradigm in the treatment of nerve injuries. Their current level of adoption among surgeons is unknown. This study evaluates the incidence of nerve transfers on case logs of board-eligible plastic surgeons over the past 14 years and surveys practicing nerve surgeons regarding their use of this technique. Methods: We queried the American Board of Plastic Surgery case log database for all nerve reconstruction Current Procedural Terminology codes from 2008 to 2021 and assessed trends and relationships between geographic region, examination year, and nerve transfer use. We surveyed nerve surgery professional societies to assess trends in practice, compared with a 2017 survey. Results: A total of 1959 nerve reconstruction cases were logged by 738 candidates from 2008 to 2021. Twelve percent of cases included nerve transfers. The proportion of nerve transfer codes (Z = −11.57; P <.0001) and the proportion of candidates performing nerve transfers (Z = −9.21, P <.0001) increased over the study period. Nerve transfers were associated with geographic region (χ2 = 25.826, P =.0002), with most cases performed in the Midwest (26.4%). A higher proportion of practicing nerve surgeons reported performing nerve transfers in this survey than in our 2017 survey (χ2 = 16.7, P <.001). Conclusions: There has been an increase in nerve transfers logged in the past 14 years by board-eligible plastic surgeons, as well as increased use among currently practicing nerve surgeons. Although nerve transfer use is increasing among both plastic and orthopedic surgeons, a greater proportion of nerve reconstructions include nerve transfers in the plastic surgery cohort. [ABSTRACT FROM AUTHOR]

In autonomous driving, environmental perception technology often encounters challenges such as false positives, missed detections, and low accuracy, particularly in detecting small objects and complex scenarios. Existing algorithms frequently suffer from issues like feature redundancy, insufficient contextual interaction, and inadequate information fusion, making it difficult to perform multi-task detection and segmentation efficiently. To address these challenges, this paper proposes an end-to-end multi-task environmental perception model named YOLO-Mg, designed to simultaneously perform traffic object detection, lane line detection, and drivable area segmentation. First, a multi-stage gated aggregation network (MogaNet) is employed during the feature extraction process to enhance contextual interaction by improving diversity in the channel dimension, thereby compensating for the limitations of feed-forward neural networks in contextual understanding. Second, to further improve the model's accuracy in detecting objects of various scales, a restructured weighted bidirectional feature pyramid network (BiFPN) is introduced, optimizing cross-level information fusion and enabling the model to handle object detection at different scales more accurately. Finally, the model is equipped with one detection head and two segmentation heads to achieve efficient multi-task environmental perception, ensuring the simultaneous execution of multiple tasks. The experimental results on the BDD100K dataset demonstrate that the model achieves a mean average precision (mAP50) of 81.4% in object detection, an Intersection over Union (IoU) of 28.9% in lane detection, and a mean Intersection over Union (mIoU) of 92.6% in drivable area segmentation. The tests conducted in real-world scenarios show that the model performs effectively, significantly enhancing environmental perception in autonomous driving and laying a solid foundation for safer and more reliable autonomous driving systems. [ABSTRACT FROM AUTHOR]

Abstract Drone aerial imaging has become increasingly important across numerous fields as drone optical sensor technology continues to advance. One critical challenge in this domain is achieving both accurate and efficient multi-object tracking. Traditional deep learning methods often separate object identification from tracking, leading to increased complexity and potential performance degradation. Conventional approaches rely heavily on manual feature engineering and intricate algorithms, which can further limit efficiency. To overcome these limitations, we propose a novel Transformer-based end-to-end multi-object tracking framework. This innovative method leverages self-attention mechanisms to capture complex inter-object relationships, seamlessly integrating object detection and tracking into a unified process. By utilizing end-to-end training, our approach simplifies the tracking pipeline, leading to significant performance improvements. A key innovation in our system is the introduction of a trajectory detection label matching technique. This technique assigns labels based on a comprehensive assessment of object appearance, spatial characteristics, and Gaussian features, ensuring more precise and logical label assignments. Additionally, we incorporate cross-frame self-attention mechanisms to extract long-term object properties, providing robust information for stable and consistent tracking. We further enhance tracking performance through a newly developed self-characteristics module, which extracts semantic features from trajectory information across both current and previous frames. This module ensures that the long-term interaction modules maintain semantic consistency, allowing for more accurate and continuous tracking over time. The refined data and stored trajectories are then used as input for subsequent frame processing, creating a feedback loop that sustains tracking accuracy. Extensive experiments conducted on the VisDrone and UAVDT datasets demonstrate the superior performance of our approach in drone-based multi-object tracking.

End-to-end systems have been introduced to address the issue of independent signal processing module design in traditional communication systems, which prevents achieving global system optimization. However, research on indoor end-to-end Visible Light Communication (VLC) systems remains limited, especially regarding the channel modeling of high-speed, high-capacity Direct Current-biased Optical Orthogonal Frequency Division Multiplexing (DCO-OFDM) systems. This paper proposes three channel modeling methods for end-to-end DCO-OFDM VLC systems. The accuracy of the proposed methods is demonstrated through R-Square model fitting performance and data distribution analysis. The effectiveness of the proposed channel modeling methods is further validated by comparing the bit error rate (BER) performance of traditional receivers and existing deep learning (DL)-based receivers. The results show that the proposed methods can effectively mitigate both linear and nonlinear distortions. By employing these channel modeling methods, communication systems can reduce the spectral occupancy of pilot signals, thereby significantly lowering the complexity of traditional channel estimation methods. Thus, these methods are suitable for use in end-to-end VLC communication systems. [ABSTRACT FROM AUTHOR]

Purpose: The current mainstream methods for single-channel speech separation generally use a feature extraction process like the short-time Fourier transform and rely on long input sequences. Thus, they do not fully utilize the information of speech features and cause signal delays in speech separation. Methods: To achieve better performance with a lightweight model, a fully convolution end-to-end audio separation network is proposed based on the features of two domains, i.e. temporal domain channel domain. It considers not only the temporal correlation of speech signals, but also the correlation between channels in the signal feature map. At first, the end-to-end network uses a convolution process with no overlapping segments to sample and encode the speech waveform. Subsequently, it calculates the mask by convolving the encoded feature space in both time series and inter-channel dimensions. Finally, it decodes the masked feature space to restructure the waveform. Results: The proposed end-to-end speech separation method makes full use of the feature space information of speech signals. Meanwhile, the separation module introduces residual structure and dilation convolution, which improves separation accuracy and computational speed with fewer parameters. The experiments show that compared with the base Conv-TasNet, the proposed model improves the SI-SNR (scale-invariant source-to-noise ratio) metric by 3.1 dB on the WSJ0-Mix2 dataset. Conclusion: This paper proposes an improved speech separation algorithm. Compared with Conv-TasNet, the performance of speech separation is improved. At the same time, the algorithm inherits the lightweight property of Conv-TasNet. In the task of separating speech signals mixed with a random signal-to-noise ratio (SNR) between −5 and 5 dB, the proposed algorithm achieves a relatively high accuracy. [ABSTRACT FROM AUTHOR]

Depth estimation represents a prevalent research focus within the realm of computer vision. Existing depth estimation methodologies utilizing LiDAR (Light Detection and Ranging) technology typically obtain sparse depth data and are associated with elevated hardware expenses. Multi-view image-matching techniques necessitate prior knowledge of camera intrinsic parameters and frequently encounter challenges such as depth inconsistency, loss of details, and the blurring of edges. To tackle these challenges, the present study introduces a monocular depth estimation approach based on an end-to-end convolutional neural network. Specifically, a DNET backbone has been developed, incorporating dilated convolution and feature fusion mechanisms within the network architecture. By integrating semantic information from various receptive fields and levels, the model's capacity for feature extraction is augmented, thereby enhancing its sensitivity to nuanced depth variations within the image. Furthermore, we introduce a loss function optimization algorithm specifically designed to address class imbalance, thereby enhancing the overall predictive accuracy of the model. Training and validation conducted on the NYU Depth-v2 (New York University Depth Dataset Version 2) and KITTI (Karlsruhe Institute of Technology and Toyota Technological Institute) datasets demonstrate that our approach outperforms other algorithms in terms of various evaluation metrics. [ABSTRACT FROM AUTHOR]

Background: The use of high-frequency electric welding technology for intestinal end-to-end anastomosis holds significant promise. Past studies have focused on in vitro, and the safety and efficacy of this technology is uncertain, severely limiting the clinical application of this technology. This study investigates the impact of compression pressure, energy dosage, and duration on anastomotic quality using a homemade anastomosis device in both in vitro and in vivo settings. Methods: Two hundred eighty intestines and 5 experimental pigs were used for in vitro and in vivo experiments, respectively. The in vitro experiments were conducted to study the effects of initial pressure (50–400 kpa), voltage (40–60 V), and time (10–20 s) on burst pressure, breaking strength, thermal damage, and histopathological microstructure of the anastomosis. Optimal parameters were then inlaid into a homemade anastomosis and used for in vivo experiments to study the postoperative porcine survival rate and the pathological structure of the tissues at the anastomosis and the characteristics of the collagen fibers. Results: The anastomotic strength was highest when the compression pressure was 250 kPa, the voltage was 60 V, and the time was 15 s. The degree of thermal damage to the surrounding tissues was the lowest. The experimental pigs had no adverse reactions after the operation, and the survival rate was 100%. 30 days after the operation, the surgical site healed well, and the tissues at the anastomosis changed from immediate adhesions to permanent connections. Conclusion: High-frequency electric welding technology has a certain degree of safety and effectiveness. It has the potential to replace the stapler anastomosis in future and become the next generation of new anastomosis device. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Nanchang University (Engineering & Technology) is the property of Nanchang University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In this paper, we introduce a novel technology utilizing relation networks with metalearning for intelligent machinery fault diagnosis, particularly in scenarios with limited labeled samples, employing the principles of few-shot learning (FSL). The proposed approach was characterized by its flexibility, simplicity, and a versatile framework. FSL facilitated the recognition and classification of new classes, requiring only a small number of samples from each category. The core of this method was the adaptive end-to-end training of the relation network (RN) from scratch. During the metalearning stage, the RN learned a deep distance metric, enabling the comparison of limited fault samples within episodes, and these episodes were simulated within the context of few-shot settings. Following the training process, the RN demonstrated the capability to classify samples from new classes by computing relation scores. Notably, it could also compare query samples with the limited samples from each new class without the need for further network updates. Experimental verification solidified the effectiveness of the proposed RN method, showcasing its robust classification abilities and achieving a relatively high level of accuracy. This technology holds promise for enhancing fault diagnosis in intelligent machinery, particularly in scenarios where labeled samples are scarce. [ABSTRACT FROM AUTHOR]

Fundamental speech recognition technologies for high-resourced languages are currently successful to build high-quality applications with the use of deep learning models. However, the problem of "borrowing" these speech recognition technologies for under-resourced languages like Vietnamese still has challenges. This study reviews fundamental studies on speech recognition in general as well as speech recognition in Vietnamese, an under-resourced language in particular. Then, it specifies the urgent issues that need current research attention to build Vietnamese speech recognition applications in practice, especially the need to build an open large sentence-labeled speech corpus and open platform for related research, which mostly benefits small individuals/organizations who do not have enough resources. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Signal Processing is the property of Journal of Signal Processing and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In current end-to-end environmental sound classification model, fixed-size filters are difficult to balance the time-frequency resolution while the weight setting of each scale is hard to reflect their importances when using multi-scale filters. Therefore, an end-to-end environmental sound classification method based on multi-scale attention residual network is proposed in this paper, which make full use of attention mechanism, muti-scale fusion and residual network structure. A weighted fusion of features at different scales by attention mechanism is utilized for better feature representation. Meanwhile, the residual structure, instead of the normal one-dimensional convolution layer, is taken into account, which alleviates the problems of gradient explosion and gradient disappearance, and accelerates the model training process. Experiments on the environmental sound datasets ESC-10, ESC-50 and UrbanSound8k show that our MSARN model achieves a classification accuracy of 91.9, 79.4 and 95.4%, respectively, which is better than other mainstream end-to-end methods. Compared with the single scale model, the accuracy is improved from 8.9 to 15.2%. [ABSTRACT FROM AUTHOR]

We present a 20-year-old patient with subglottic and tracheal stenosis was taken for a tracheal resection and end-to-end anastomosis. The patient's neck was positioned in hyperflexion using chin stitches to minimize tension at the anastomosis. On post-operative period, the patient developed paresthesias in upper and lower extremities associated with motor weakness. Magnetic resonance imaging was performed showing lesions compromising ventral spinal cord at the level of C4–C5 and C6–C7. Chin stitches were removed and neck flexion was reduced. The patient remained in the intensive care unit with vasopressors, physical therapy and intravenous fluid-therapy to maintain mean arterial pressure above 90 mmHg. After 3 weeks, the patient was discharged with no neurologic deficit. There are few cases reported of acute ischemic spinal injury following tracheal reconstruction. If this complication arises, neck posture should be corrected, maintenance of MAP above 90 mmHg and implementation of early physical therapy is key to improve neurologic outcomes. [ABSTRACT FROM AUTHOR]

In autonomous driving, environmental perception technology often encounters challenges such as false positives, missed detections, and low accuracy, particularly in detecting small objects and complex scenarios. Existing algorithms frequently suffer from issues like feature redundancy, insufficient contextual interaction, and inadequate information fusion, making it difficult to perform multi-task detection and segmentation efficiently. To address these challenges, this paper proposes an end-to-end multi-task environmental perception model named YOLO-Mg, designed to simultaneously perform traffic object detection, lane line detection, and drivable area segmentation. First, a multi-stage gated aggregation network (MogaNet) is employed during the feature extraction process to enhance contextual interaction by improving diversity in the channel dimension, thereby compensating for the limitations of feed-forward neural networks in contextual understanding. Second, to further improve the model’s accuracy in detecting objects of various scales, a restructured weighted bidirectional feature pyramid network (BiFPN) is introduced, optimizing cross-level information fusion and enabling the model to handle object detection at different scales more accurately. Finally, the model is equipped with one detection head and two segmentation heads to achieve efficient multi-task environmental perception, ensuring the simultaneous execution of multiple tasks. The experimental results on the BDD100K dataset demonstrate that the model achieves a mean average precision (mAP50) of 81.4% in object detection, an Intersection over Union (IoU) of 28.9% in lane detection, and a mean Intersection over Union (mIoU) of 92.6% in drivable area segmentation. The tests conducted in real-world scenarios show that the model performs effectively, significantly enhancing environmental perception in autonomous driving and laying a solid foundation for safer and more reliable autonomous driving systems.

End-to-end systems have been introduced to address the issue of independent signal processing module design in traditional communication systems, which prevents achieving global system optimization. However, research on indoor end-to-end Visible Light Communication (VLC) systems remains limited, especially regarding the channel modeling of high-speed, high-capacity Direct Current-biased Optical Orthogonal Frequency Division Multiplexing (DCO-OFDM) systems. This paper proposes three channel modeling methods for end-to-end DCO-OFDM VLC systems. The accuracy of the proposed methods is demonstrated through R-Square model fitting performance and data distribution analysis. The effectiveness of the proposed channel modeling methods is further validated by comparing the bit error rate (BER) performance of traditional receivers and existing deep learning (DL)-based receivers. The results show that the proposed methods can effectively mitigate both linear and nonlinear distortions. By employing these channel modeling methods, communication systems can reduce the spectral occupancy of pilot signals, thereby significantly lowering the complexity of traditional channel estimation methods. Thus, these methods are suitable for use in end-to-end VLC communication systems.

Depth estimation represents a prevalent research focus within the realm of computer vision. Existing depth estimation methodologies utilizing LiDAR (Light Detection and Ranging) technology typically obtain sparse depth data and are associated with elevated hardware expenses. Multi-view image-matching techniques necessitate prior knowledge of camera intrinsic parameters and frequently encounter challenges such as depth inconsistency, loss of details, and the blurring of edges. To tackle these challenges, the present study introduces a monocular depth estimation approach based on an end-to-end convolutional neural network. Specifically, a DNET backbone has been developed, incorporating dilated convolution and feature fusion mechanisms within the network architecture. By integrating semantic information from various receptive fields and levels, the model’s capacity for feature extraction is augmented, thereby enhancing its sensitivity to nuanced depth variations within the image. Furthermore, we introduce a loss function optimization algorithm specifically designed to address class imbalance, thereby enhancing the overall predictive accuracy of the model. Training and validation conducted on the NYU Depth-v2 (New York University Depth Dataset Version 2) and KITTI (Karlsruhe Institute of Technology and Toyota Technological Institute) datasets demonstrate that our approach outperforms other algorithms in terms of various evaluation metrics.

This study investigates the utilization of complex operations to perform multichannel speech enhancement in the time domain using a neural network. Previous studies have demonstrated the advantages of incorporating complex operations when designing neural networks; however, they have solely focused on frequency-domain enhancement techniques. In contrast, our research study presents an end-to-end approach to perform speech enhancement in the time domain. We used the Hilbert transform to intelligently generate complex time-domain waveforms as inputs to the network. This allowed us to create an end-to-end approach that explores spatial information. To handle the complexity of the inputs, we developed a complex neural adaptive beamformer (CNAB). We utilized complex shared long short-term memory (LSTM), split-LSTM, and complex convolutions to generate the beamforming output. Following this, we developed a complex full convolutional network (CFCN) to enhance the beamforming output. We leveraged complex dilated convolutions to model the long-term temporal dependencies of speech. By cascading the CNAB and CFCN, we created the final end-to-end time-domain enhancement network, named CNABCFCN. We trained and tested CNABCFCN using the deep noise suppression (DNS) challenge dataset. Our results demonstrate the advantages of using complex operations over the baseline model. Furthermore, the proposed CNABCFCN performed better in terms of both objective and subjective measures compared with other networks. [ABSTRACT FROM AUTHOR]

Object detection from high-resolution remote sensing (RS) imagery plays an important role in a wide range of applications, including traffic management, urban planning, and automatic surveying and mapping. Meanwhile, the end-to-end model for object detection is becoming a hot topic in computer vision area. However, owing to the complex background in the RS image, the results of object detection based on the end-to-end model are not satisfactory. Particularly, the issue lies in the small and large objects, which are easily omitted. We introduce advanced end-to-end object detection with transformers with improved deNoising anchOr box (DINO) and propose a new framework improvement for RS image object detection, called improved detection method for full-scale objects based on DINO. First, we propose a feature-rich hybrid stacking module. To retain more multiscale features during channel mapping, we add two additional mapping superposition channels to conventional channel mapping. Second, we propose a low-level attention reuse module and believe that the information in the low-level attention model is incomplete and needs to be reused. Therefore, we add the output of low-level attention to the input of next-level attention through a simple activation module. We conducted experimental verification on the AIcrowd dataset and the DOTA-v1.0 dataset. On the AIcrowd and DOTA-v1.0 datasets, compared with the mainstream algorithms, our algorithm has improved the accuracy of full-scale objects, especially for large objects. In addition, our model surpasses several state-of-the-art methods, including recent anchor-free methods, in terms of accuracy on the DOTA-v1.0 dataset.

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In this paper, we propose an end-to-end cross-layer gated attention network (CLGA-Net) to directly restore fog-free images. Compared with the previous dehazing network, the dehazing model presented in this paper uses the smooth cavity convolution and local residual module as the feature extractor, combined with the channel attention mechanism, to better extract the restored features. A large amount of experimental data proves that the defogging model proposed in this paper is superior to previous defogging technologies in terms of structure similarity index (SSIM), peak signal to noise ratio (PSNR) and subjective visual quality. In order to improve the efficiency of decoding and encoding,we also describe a fusion residual module and conduct ablation experiments, which prove that the fusion residual is suitable for the dehazing problem. Therefore, we use fusion residual as a fixed module for encoding and decoding. In addition, we found that the traditional defogging model based on the U-net network may cause some information losses in space. We have achieved effective maintenance of low-level feature information through the cross-layer gating structure that better takes into account global and subtle features. We also present the application of our CLGA-Net in challenging scenarios where the best results in both quantity and quality can be obtained. Experimental results indicate that the present cross-layer gating module can be widely used in the same type of network. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Central China Normal University is the property of Huazhong Normal University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Aspect-based sentiment analysis (ABSA) aims to conduct fine-grained sentiment analysis, necessitating the extraction of three key components: target entity, aspect category and sentiment polarity. These three components collectively form an integrated ABSA task known as TASD (Target-Aspect-Sentiment jointly Detection). Most of existing approaches on ABSA usually employ Recurrent neural networks(RNNs), Convolutional neural networks(CNNs) or pre-training models such as Bidirectional Encoder Representations from Transformers(BERT). However, they have some common weaknesses. First, most of the existing methods focus on one or two sub-tasks instead of triplet detection, thus they do not establish an end-to-end (training a complex learning system represented by a single model) ABSA model and cannot utilize the relevance of multiple ABSA sub-tasks during training. Second, they cannot achieve accuracy and efficiency simultaneously due to the coupling of context and given aspects. Third, they are poor in recognizing implicit targets. To tackle these limitations, this paper proposes a novel method named the Twin Towers End to End model (TTEE) to solve TASD task. It transforms complex TASD task into a simple end-to-end multi-task framework, simultaneously conducting target and aspect-sentiment detection. It builds twin towers system based on BERT or its updated versions to decouple context and given aspects, which can reduce redundant calculation to improve computational efficiency significantly. It offers great advantage to identify implicit target entity and its associated aspect-sentiment in the context without introducing extra model architecture. Experiments on three public datasets in different domains demonstrate that our approach not only achieves better performance on various evaluation metrics, but also has high efficiency in both training and inference phases, over a wide range of sample size and number of aspect categories. • The T-A-S triplets can be decoded from the end-to-end output of TTEE. • BERT-based model is employed to encode context and aspect-sentiment pairs separately. • TTEE efficiently reduces redundant calculations of embedding process. • TTEE excels in detecting implicit targets within the context. [ABSTRACT FROM AUTHOR]

Traditionally, teaching a human and a Machine Learning (ML) model is quite different, but organized and structured learning has the ability to enable faster and better understanding of the underlying concepts. For example, when humans learn to speak, they first learn how to utter basic phones and then slowly move towards more complex structures such as words and sentences. Motivated by this observation, researchers have started to adapt this approach for training ML models. Since the main concept, the gradual increase in difficulty, resembles the notion of the curriculum in education, the methodology became known as Curriculum Learning (CL). In this work, we design and test new CL approaches to train Automatic Speech Recognition systems, specifically focusing on the so-called end-to-end models. These models consist of a single, large-scale neural network that performs the recognition task, in contrast to the traditional way of having several specialized components focusing on different subtasks (e.g., acoustic and language modeling). We demonstrate that end-to-end models can achieve better performances if they are provided with an organized training set consisting of examples that exhibit an increasing level of difficulty. To impose structure on the training set and to define the notion of an easy example, we explored multiple solutions that use either external, static scoring methods or incorporate feedback from the model itself. In addition, we examined the effect of pacing functions that control how much data is presented to the network during each training epoch. Our proposed curriculum learning strategies were tested on the task of speech recognition on two data sets, one containing spontaneous Finnish speech where volunteers were asked to speak about a given topic, and one containing planned English speech. Empirical results showed that a good curriculum strategy can yield performance improvements and speed-up convergence. After a given number of epochs, our best strategy achieved a 5.6% and 3.4% decrease in terms of test set word error rate for the Finnish and English data sets, respectively. • Curriculum learning can improve the performance of end-to-end ASR systems. • A good curriculum strategy can speed up convergence without impacting performance. • Curriculum learning shows benefits in both low and high resource settings. • An adaptive curriculum leads to faster convergence rates than a static curriculum. [ABSTRACT FROM AUTHOR]

Recent advances in deep learning have contributed greatly to the field of parallel MR imaging, where a reduced amount of k-space data are acquired to accelerate imaging time. In our previous work, we have proposed a deep learning method to reconstruct MR images directly from k-space data acquired with Cartesian trajectories. However, MRI utilizes various non-Cartesian trajectories, such as radial trajectories, with various numbers of multi-channel RF coils according to the purpose of an MRI scan. Thus, it is important for a reconstruction network to efficiently unfold aliasing artifacts due to undersampling and to combine multi-channel k-space data into single-channel data. In this work, a neural network named 'ETER-net' is utilized to reconstruct an MR image directly from k-space data acquired with Cartesian and non-Cartesian trajectories and multi-channel RF coils. In the proposed image reconstruction network, the domain transform network converts k-space data into a rough image, which is then refined in the following network to reconstruct a final image. We also analyze loss functions including adversarial and perceptual losses to improve the network performance. For experiments, we acquired k-space data at a 3T MRI scanner with Cartesian and radial trajectories to show the learning mechanism of the direct mapping relationship between the k-space and the corresponding image by the proposed network and to demonstrate the practical applications. According to our experiments, the proposed method showed satisfactory performance in reconstructing images from undersampled single- or multi-channel k-space data with reduced image artifacts. In conclusion, the proposed method is a deep-learning-based MR reconstruction network, which can be used as a unified solution for parallel MRI, where k-space data are acquired with various scanning trajectories. [ABSTRACT FROM AUTHOR]

We propose a conceptually simple, general framework and end-to-end approach to point cloud completion, entitled PCA-Net. This approach differs from the existing methods in that it does not require a "simple" network, such as multilayer perceptrons (MLPs), to generate a coarse point cloud and then a "complex" network, such as auto-encoders or transformers, to enhance local details. It can directly learn the mapping between missing and complete points, ensuring that the structure of the input missing point cloud remains unchanged while accurately predicting the complete points. This approach follows the minimalist design of U-Net. In the encoder, we encode the point clouds into point cloud blocks by iterative farthest point sampling (IFPS) and k-nearest neighbors and then extract the depth interaction features between the missing point cloud blocks by the attention mechanism. In the decoder, we introduce a new trilinear interpolation method to recover point cloud details, with the help of the coordinate space and feature space of low-resolution point clouds, and missing point cloud information. This paper also proposes a method to generate multi-view missing point cloud data using a 3D point cloud hidden point removal algorithm, so that each 3D point cloud model generates a missing point cloud through eight uniformly distributed camera poses. Experiments validate the effectiveness and superiority of PCA-Net in several challenging point cloud completion tasks, and PCA-Net also shows great versatility and robustness in real-world missing point cloud completion. [ABSTRACT FROM AUTHOR]

Drone aerial imaging has become increasingly important across numerous fields as drone optical sensor technology continues to advance. One critical challenge in this domain is achieving both accurate and efficient multi-object tracking. Traditional deep learning methods often separate object identification from tracking, leading to increased complexity and potential performance degradation. Conventional approaches rely heavily on manual feature engineering and intricate algorithms, which can further limit efficiency. To overcome these limitations, we propose a novel Transformer-based end-to-end multi-object tracking framework. This innovative method leverages self-attention mechanisms to capture complex inter-object relationships, seamlessly integrating object detection and tracking into a unified process. By utilizing end-to-end training, our approach simplifies the tracking pipeline, leading to significant performance improvements. A key innovation in our system is the introduction of a trajectory detection label matching technique. This technique assigns labels based on a comprehensive assessment of object appearance, spatial characteristics, and Gaussian features, ensuring more precise and logical label assignments. Additionally, we incorporate cross-frame self-attention mechanisms to extract long-term object properties, providing robust information for stable and consistent tracking. We further enhance tracking performance through a newly developed self-characteristics module, which extracts semantic features from trajectory information across both current and previous frames. This module ensures that the long-term interaction modules maintain semantic consistency, allowing for more accurate and continuous tracking over time. The refined data and stored trajectories are then used as input for subsequent frame processing, creating a feedback loop that sustains tracking accuracy. Extensive experiments conducted on the VisDrone and UAVDT datasets demonstrate the superior performance of our approach in drone-based multi-object tracking., (© 2024. The Author(s).)

This study investigates the utilization of complex operations to perform multichannel speech enhancement in the time domain using a neural network. Previous studies have demonstrated the advantages of incorporating complex operations when designing neural networks; however, they have solely focused on frequency-domain enhancement techniques. In contrast, our research study presents an end-to-end approach to perform speech enhancement in the time domain. We used the Hilbert transform to intelligently generate complex time-domain waveforms as inputs to the network. This allowed us to create an end-to-end approach that explores spatial information. To handle the complexity of the inputs, we developed a complex neural adaptive beamformer (CNAB). We utilized complex shared long short-term memory (LSTM), split-LSTM, and complex convolutions to generate the beamforming output. Following this, we developed a complex full convolutional network (CFCN) to enhance the beamforming output. We leveraged complex dilated convolutions to model the long-term temporal dependencies of speech. By cascading the CNAB and CFCN, we created the final end-to-end time-domain enhancement network, named CNABCFCN. We trained and tested CNABCFCN using the deep noise suppression (DNS) challenge dataset. Our results demonstrate the advantages of using complex operations over the baseline model. Furthermore, the proposed CNABCFCN performed better in terms of both objective and subjective measures compared with other networks.

Simple Summary: Trunk-boring insects belong to one of the most destructive forest pests. Larvae in some groups are particularly difficult to detect since they make their living in trunks and no obvious sign can be found from outside. To deal with this problem, a new method is presented here, which is the embedding of a vibration probe into the tree trunk to pick up vibrations caused by larvae and the use of a model to distinguish whether the tree is infected. However, this procedure can experience severe interference from environmental noise, which is simultaneously picked up with vibrations. Thus, it is necessary to add a noise suppression process before discrimination. Previous examples have proved that the application of the analysis intended for sounds to boring vibrations is feasible. Therefore, we took advantage of deep learning-based speech enhancement and further improved it to develop a boring vibration enhancement model. The training data used in this research contains boring vibrations recorded within pieces of trunks and noise, which are common for trees' living environment. The experimental results indicate that the enhancement procedure provided by our model substantially increases the accuracy of several well-known classification models, guaranteeing a more practical larvae detection. The larvae of some trunk-boring beetles barely leave traces on the outside of trunks when feeding within, rendering the detection of them rather difficult. One approach to solving this problem involves the use of a probe to pick up boring vibrations inside the trunk and distinguish larvae activity according to the vibrations. Clean boring vibration signals without noise are critical for accurate judgement. Unfortunately, these environments are filled with natural or artificial noise. To address this issue, we constructed a boring vibration enhancement model named VibDenoiser, which makes a significant contribution to this rarely studied domain. This model is built using the technology of deep learning-based speech enhancement. It consists of convolutional encoder and decoder layers with skip connections, and two layers of SRU++ for sequence modeling. The dataset constructed for study is made up of boring vibrations of Agrilus planipennis Fairmaire, 1888 (Coleoptera: Buprestidae) and environmental noise. Our VibDenoiser achieves an improvement of 18.57 in SNR, and it runs in real-time on a laptop CPU. The accuracy of the four classification models increased by a large margin using vibration clips enhanced by our model. The results demonstrate the great enhancement performance of our model, and the contribution of our work to better boring vibration detection. [ABSTRACT FROM AUTHOR]

Around the world, the video surveillance system has gained wide acceptance and astonishing growth due to its broad applications. The surveillance system has become a paramount tool and benchmark for analyzing the harmony and safety of society. Anomaly detection and its associated applications play a key role in the integrity of the system. The aim of anomaly detection is to find rare and sparse occurrences of events from videos. Developing an accurate and time-efficient system is still remains challenging due to the dynamic nature of a nomalies. The deep learning-based end-to-end system with full use of both spatial and temporal features from the input videos is proposed. The model combines the use of 2DCNN and Stacked LSTM to extract frame-level features through an anisotropic Gunnar Farneback Optical Flow algorithm. The system is evaluated on the benchmarked datasets namely UCSD Ped1 and UCSD Ped2, and it achieves an AUC of 95% and 94% respectively. The experimental results indicate that the proposed method is superior to state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

Rock discontinuities are essential for the mechanical behavior and stability of rock mass. Previous approaches for characterizing discontinuities either rely on limited handcrafted features (point normals, point curvatures, point densities, and so on) or fail to classify discontinuities, making them unsuitable for complex and large-scale scenes. To cope with these problems, an end-to-end point-based deep learning method that can automatically learn rich and high-dimensional features and classify discontinuities was developed in this study. Firstly, a roadcut and part of a natural slope were selected to train the developed network and assess its performance. Subsequently, the trained network was used to classify the remaining part of the slope. Finally, the "Density-Based Scan Algorithm with Noise" (DBSCAN) and principal component analysis (PCA) algorithms were employed to extract individual discontinuities and calculate their orientations. The two cases achieved a global accuracy (GA) of 97.25% and 94.56%, respectively, and a mean intersection over union (MIoU) of 93.77% and 88.66%, respectively, indicating the excellent performance of the network. Meanwhile, the average error in dip angle and dip direction was 0.67° and 3.33°, respectively, proving the characterization ability of the developed method was satisfactory. Furthermore, the presented method exhibits strong robustness and the potential to characterize large-scale rock discontinuities with noise. This method facilitates the application of deep learning in geosciences and provides geologists and geological engineers with a new idea for rock discontinuity characterization. • A point-based deep learning method for discontinuity characterization was developed. • The method can process point cloud directly without handcrafted features. • It can extract rich and high-dimensional features and classify discontinuities. • It has huge advantages in mapping discontinuities in complex and large-scale scenes. [ABSTRACT FROM AUTHOR]

Introduction: Two end‐to‐end tests evaluate the accuracy of a surface‐guided radiation therapy (SGRT) system (CRAD Catalyst HD) for position verification in comparison to a stereoscopic x‐ray imaging system (Brainlab Exactrac) for single‐isocenter, multiple metastases stereotactic radiosurgery (SRS) using 3D polymer gel inserts. Materials and methods: A 3D‐printed phantom (Prime phantom, RTsafe PC, Athens, Greece) with two separate cylindrical polymer gel inserts were immobilized in open‐face masks and treated with a single isocentric, multitarget SRS plan. Planning was done in Brainlab (Elements) to treat five metastatic lesions in one fraction, and initial setup was done using cone beam computed tomography. Positional verification was done using orthogonal X‐ray imaging (Brainlab Exactrac) and/or a surface imaging system (CRAD Catalyst HD, Uppsala, Sweden), and shift discrepancies were recorded for each couch angle. Forty‐two hours after irradiation, the gel phantom was scanned in a 1.5 Tesla MRI, and images were fused with the patient computed tomography data/structure set for further analysis of spatial dose distribution. Results: Discrepancies between the CRAD Catalyst HD system and Brainlab Exactrac were <1 mm in the translational direction and <0.5° in the angular direction at noncoplanar couch angles. Dose parameters (DMean%,D95%) and 3D gamma index passing rates were evaluated for both setup modalities for each planned target volume (PTV) at a variety of thresholds: 3%/2 mm (Exactrac≥93.1% and CRAD ≥87.2%), 5%/2 mm (Exactrac≥95.6% and CRAD ≥94.6%), and 5%/1 mm (Exactrac≥81.8% and CRAD ≥83.7%). Conclusion: Dose metrics for a setup with surface imaging was found to be consistent with setup using x‐ray imaging, demonstrating high accuracy and reproducibility for treatment delivery. Results indicate the feasibility of using surface imaging for position verification at noncoplanar couch angles for single‐isocenter, multiple‐target SRS using end‐to‐end quality assurance (QA) testing with 3D polymer gel dosimetry. [ABSTRACT FROM AUTHOR]

Dialogue system aims to realize the communication for information between machine and human so as to provide personalized service, it has certain research value. In order to follow up the research in related fields timely, this paper reviewed the current research progress. Firstly, it introduced the development and classification of dialogue system. Secondly, it summarized the architecture of task-oriented dialogue system, focusing on module based and the end-to-end based frameworks, then summarized the advantages and disadvantages of the models. Thirdly, it introduced the evaluation methods and application fields of task-oriented dialogue system, including e-commerce, education, medicine and so on. Finally, it analyzed and summarized the problems and challenges faced by the multi-turn task-oriented dialogue system. [ABSTRACT FROM AUTHOR]

Human action recognition is a challenging task in computer vision. Modeling spatial-temporal information in videos effectively is crucial for the performance improvement of action recognition. In this paper, we introduce an end to end framework, Spatio-Temporal Attention ConvNet (STACNet), which combines two novel attention modules and convolutional neural networks for action recognition. Two novel attention modules, Spatial Attention Module (SAM) and Temporal Attention Module (TAM), are proposed respectively. Spatial Attention Module is established by fusing the value feature and the gradient feature of the feature map, making the representation of ConvNet for action recognition focus on the informative motion regions of actions. Temporal Attention Module is built by combining global average pooling and global max pooling to explore key frames in videos. With the two attention modules, STACNet can adaptively distinguish key frames in a sequence and selectively pay different levels of attention to different spatial motion regions of human actions, at virtually negligible increase in computation cost. We demonstrate the effectiveness of SAM and TAM for action recognition, respectively. The experimental results show that STACNet can obtain superior performance on the datasets of HMDB51 and UCF101.

The aim of the study is to investigate the possibility of reinnervation and prevention of muscle atrophy following side to side neurorrhaphy to the intact nerve after end to end repair of proximal transected peripheral nerves in order to prevent distal target organ atrophy. For this, four groups each containing five SpragueDawley female rats were used. In group I, no surgical procedure was performed as a control group. In group II, side to side distal neurorrhaphy performed to the tibial and peroneal nerves after end to end repair of transected proximal tibial nerve. In group III. distally side to side epineural neurorrhaphy performed to the tibial and peroneal nerves. In group IV, end to end epineural repair was performed after proximal tibial nerve transection. The rats were followed up for 3 months for nerve regeneration. Subsequently group II, III and IV were evaluated histopathologically. In all group, tibial and fibular bony weights, foot weights, anterior and posterior crural muscle weights and EMG parameters were evaluated. Comparison between the groups revealed no significant differences regarding EMG and muscle weights between groups 2 and 4 also axonal degeneration was observed in 3 group after neurorrhaphy. As result of experimental study, we think that side to side repair of intact distal nerves as an adjunct to end to end repair of proximal nerve transections has no additional benefit to prevent distal organ atrophy but rather may be caused harm on the intact nerve. In addition, it has been observed that this technique affects the intact nerve rather than the transected nerve. [Med-Science 2019; 8(4.000): 827-31]

Performance of Automatic Speech Recognition (ASR) systems is known to suffer considerable degradation when exposed to Far-Field speech data capture. Consequently, far-field ASR has received considerable attention in recent years. Motivated by our recent work using Curriculum Learning (CL) based strategies to improve Speaker Identification (SID) under noisy and degraded conditions, this study proposes a novel approach to improve far-field ASR using CL based approaches. Specifically, we propose using a CL based approach for training a Bidirectional Long Short Term Memory (BLSTM) based ASR network trained using the Connectionist Temporal Classification (CTC) objective function. We initiate the training with comparatively easier near-field data, and include more diverse (difficult) far-field data progressively in the later stages of training. These proposed approaches are shown to significantly outperform the baseline BLSTM ASR system, and offer relative reductions in WERs of up to +7.3% and +10.1% for the dev and eval sets of the AMI far-field voice capture corpus. [ABSTRACT FROM AUTHOR]

Recent advances in deep learning have contributed greatly to the field of parallel MR imaging, where a reduced amount of k-space data are acquired to accelerate imaging time. In our previous work, we have proposed a deep learning method to reconstruct MR images directly from k-space data acquired with Cartesian trajectories. However, MRI utilizes various non-Cartesian trajectories, such as radial trajectories, with various numbers of multi-channel RF coils according to the purpose of an MRI scan. Thus, it is important for a reconstruction network to efficiently unfold aliasing artifacts due to undersampling and to combine multi-channel k-space data into single-channel data. In this work, a neural network named ‘ETER-net’ is utilized to reconstruct an MR image directly from k-space data acquired with Cartesian and non-Cartesian trajectories and multi-channel RF coils. In the proposed image reconstruction network, the domain transform network converts k-space data into a rough image, which is then refined in the following network to reconstruct a final image. We also analyze loss functions including adversarial and perceptual losses to improve the network performance. For experiments, we acquired k-space data at a 3T MRI scanner with Cartesian and radial trajectories to show the learning mechanism of the direct mapping relationship between the k-space and the corresponding image by the proposed network and to demonstrate the practical applications. According to our experiments, the proposed method showed satisfactory performance in reconstructing images from undersampled single- or multi-channel k-space data with reduced image artifacts. In conclusion, the proposed method is a deep-learning-based MR reconstruction network, which can be used as a unified solution for parallel MRI, where k-space data are acquired with various scanning trajectories.

Objectives/Hypothesis: The literature on outcomes of end‐to‐side (ETS) anastomoses for microvascular reconstruction of the head and neck is limited. This series reviews ETS in free tissue transfer (FTT) across multiple institutions to better understand their usage and associated outcomes. Study Design: Retrospective review of 2482 consecutive patients across three tertiary institutions. Methods: Adult patients (> 18) who received a FTT from 2006 to 2019 were included. Results: Two hundred and twenty‐one FTT were identified as requiring at least one ETS anastomosis. These ETS cases had a failure rate of 11.2% in comparison to 3.8% in a cohort of end‐to‐end (ETE) cases (P <.001). ETS cases were significantly more likely to have a prior neck dissection (P <.001), suggesting the ETS method was utilized in select circumstances. A second ETS anastomosis improved survival of the FTT (P =.006), as did utilization of a coupler over suture (P =.002). Failure due to venous thrombosis was significantly more common with one ETS anastomosis instead of two ETS anastomoses (P =.042). Conclusions: ETS is effective but is often used as a secondary technique when ETE is not feasible; as such, in this series, ETS was associated with higher failure. A second anastomosis and the use of the coupler for completing the anastomoses were associated with lower rates of failure. Level of Evidence: 3 Laryngoscope, 131:1286–1290, 2021 [ABSTRACT FROM AUTHOR]

In laparoscopic cholecystectomy (LC), the anatomical variation of gallbladder canal should be noted. Bile duct injury is one of the most serious complications of LC, which might lead to serious complications. This case because of an intraoperative accidental cut to right anterior inferior lobe bile duct which joins to the cystic duct, after confirming by intraoperative magnetic resonance imaging-guided real-time three-dimensional bile duct reconstruction and performing end-to-end anastomosis of the right anterior inferior lobe bile duct and the gallbladder tube under laparoscopy, the patient was discharged 5 days after surgery, was followed up for 4 months and was disease-free.

Computational drug repositioning plays a vital role in the prediction of drug function. Many new functions discovered have been confirmed. In comparison with traditional drug repositioning, computational drug repositioning shortens the time and reduces labor. Thus, it has received wide attention in recent years. However, prediction remains a considerable challenge. In this paper, a method called HNRD is introduced to predict the link between drugs and diseases. It is based on neighborhood information aggregation in neural networks which combines the similarity of diseases and drugs, the associations between the drugs and diseases. Compared with the state-of-the-art method before, our method has achieved better results, with the best AUC of 0.97 in one of the golden datasets. To better evaluate our approach, we also performed data analysis based on one-to-one association's prediction and robust analysis by testing on different datasets. All the results prove the excellent performance of prediction. Source codes of this paper are available on https://github.com/heibaipei/HNRD.

We propose a conceptually simple, general framework and end-to-end approach to point cloud completion, entitled PCA-Net. This approach differs from the existing methods in that it does not require a “simple” network, such as multilayer perceptrons (MLPs), to generate a coarse point cloud and then a “complex” network, such as auto-encoders or transformers, to enhance local details. It can directly learn the mapping between missing and complete points, ensuring that the structure of the input missing point cloud remains unchanged while accurately predicting the complete points. This approach follows the minimalist design of U-Net. In the encoder, we encode the point clouds into point cloud blocks by iterative farthest point sampling (IFPS) and k-nearest neighbors and then extract the depth interaction features between the missing point cloud blocks by the attention mechanism. In the decoder, we introduce a new trilinear interpolation method to recover point cloud details, with the help of the coordinate space and feature space of low-resolution point clouds, and missing point cloud information. This paper also proposes a method to generate multi-view missing point cloud data using a 3D point cloud hidden point removal algorithm, so that each 3D point cloud model generates a missing point cloud through eight uniformly distributed camera poses. Experiments validate the effectiveness and superiority of PCA-Net in several challenging point cloud completion tasks, and PCA-Net also shows great versatility and robustness in real-world missing point cloud completion.

The larvae of some trunk-boring beetles barely leave traces on the outside of trunks when feeding within, rendering the detection of them rather difficult. One approach to solving this problem involves the use of a probe to pick up boring vibrations inside the trunk and distinguish larvae activity according to the vibrations. Clean boring vibration signals without noise are critical for accurate judgement. Unfortunately, these environments are filled with natural or artificial noise. To address this issue, we constructed a boring vibration enhancement model named VibDenoiser, which makes a significant contribution to this rarely studied domain. This model is built using the technology of deep learning-based speech enhancement. It consists of convolutional encoder and decoder layers with skip connections, and two layers of SRU++ for sequence modeling. The dataset constructed for study is made up of boring vibrations of Agrilus planipennis Fairmaire, 1888 (Coleoptera: Buprestidae) and environmental noise. Our VibDenoiser achieves an improvement of 18.57 in SNR, and it runs in real-time on a laptop CPU. The accuracy of the four classification models increased by a large margin using vibration clips enhanced by our model. The results demonstrate the great enhancement performance of our model, and the contribution of our work to better boring vibration detection.