Deep Learning-based Algorithm Research Articles

Impact of rapid iodine contrast agent infusion on tracheal diameter and lung volume in CT pulmonary angiography measured with deep learning-based algorithm.

To compare computed tomography (CT) pulmonary angiography and unenhanced CT to determine the effect of rapid iodine contrast agent infusion on tracheal diameter and lung volume. This retrospective study included 101 patients who underwent CT pulmonary angiography and unenhanced CT, for which the time interval between them was within 365 days. CT pulmonary angiography was scanned 20 s after starting the contrast agent injection at the end-inspiratory level. Commercial software, which was developed based on deep learning technique, was used to segment the lung, and its volume was automatically evaluated. The tracheal diameter at the thoracic inlet level was also measured. Then, the ratios for the CT pulmonary angiography to unenhanced CT of the tracheal diameter (TDPAU) and both lung volumes (BLVPAU) were calculated. Tracheal diameter and both lung volumes were significantly smaller in CT pulmonary angiography (17.2 ± 2.6mm and 3668 ± 1068ml, respectively) than those in unenhanced CT (17.7 ± 2.5mm and 3887 ± 1086ml, respectively) (p < 0.001 for both). A statistically significant correlation was found between TDPAU and BLVPAU with a correlation coefficient of 0.451 (95% confidence interval, 0.280-0.594) (p < 0.001). No factor showed a significant association with TDPAU. The type of contrast agent had a significant association for BLVPAU (p = 0.042). Rapid infusion of iodine contrast agent reduced the tracheal diameter and both lung volumes in CT pulmonary angiography, which was scanned at end-inspiratory level, compared with those in unenhanced CT.

Prediction of Mismatch Repair Status in Endometrial Cancer from Histological Slide Images Using Various Deep Learning-Based Algorithms

The application of deep learning algorithms to predict the molecular profiles of various cancers from digital images of hematoxylin and eosin (H&amp;E)-stained slides has been reported in recent years, mainly for gastric and colon cancers. In this study, we investigated the potential use of H&amp;E-stained endometrial cancer slide images to predict the associated mismatch repair (MMR) status. H&amp;E-stained slide images were collected from 127 cases of the primary lesion of endometrial cancer. After digitization using a Nanozoomer virtual slide scanner (Hamamatsu Photonics), we segmented the scanned images into 5397 tiles of 512 × 512 pixels. The MMR proteins (PMS2, MSH6) were immunohistochemically stained, classified into MMR proficient/deficient, and annotated for each case and tile. We trained several neural networks, including convolutional and attention-based networks, using tiles annotated with the MMR status. Among the tested networks, ResNet50 exhibited the highest area under the receiver operating characteristic curve (AUROC) of 0.91 for predicting the MMR status. The constructed prediction algorithm may be applicable to other molecular profiles and useful for pre-screening before implementing other, more costly genetic profiling tests.

Assessment of changes in vessel area during needle manipulation in microvascular anastomosis using a deep learning-based semantic segmentation algorithm: A pilot study

Assessment of changes in vessel area during needle manipulation in microvascular anastomosis using a deep learning-based semantic segmentation algorithm: A pilot study

Evaluation of Cell Cycle-Dependent Migration Activity after X-ray Exposure: A Radiobiological Approach for Optimization of Radiotherapy with Cell Cycle-Targeting Agents.

Radiotherapy with cell cycle-specific anticancer agents has become an important option in the control of both primary tumors and metastases. Here, we used image analysis algorithms that enable quick segmentation and tracking to describe a radiobiological approach for the optimized selection of cell cycle-targeting anticancer drugs for radiotherapy. We confirmed cell cycle-synchronization using human cervical cancer HeLa cells expressing a fluorescent ubiquitination-based cell cycle indicator (FUCCI) as a cell cycle-monitoring probe. Cells synchronized in the G1 and G2 phases were irradiated with X rays at 0.5-2 Gy. Each cell was identified using Cellpose, a deep learning-based algorithm for cellular segmentation, and the velocity and direction of migration were analyzed using the TrackMate plugin in Fiji ImageJ. G1 phase synchronized cells showed a dose-dependent decrease in velocity after irradiation, while G2 cells tended to increase their velocity. The migration pattern of all cells appeared to be a random walk model, regardless of the exposure dose. In addition, we used cisplatin to arrest the cell cycle. HeLa-FUCCI cells arrested at the G2 phase via cisplatin treatment showed enhanced cell migration after X-ray exposure. These results indicated that anticancer agents that arrest the cell cycle of cancer cells in a specific phase may enhance cell migration after radiotherapy. Our approach, using cellular segmentation and tracking algorithms, could enhance the radiobiological assessment of cell cycle-specific migration after irradiation to aid in optimizing radiotherapy using cell cycle-targeting agents.

FSRDiff: A fast diffusion-based super-resolution method using GAN

Single image super-resolution with diffusion probabilistic models (SRDiff) is a successful diffusion model for image super-resolution that produces high-quality images and is stable during training. However, due to the long sampling time, it is slower in the testing phase than other deep learning-based algorithms. Reducing the total number of diffusion steps can accelerate sampling, but it also causes the inverse diffusion process to deviate from the Gaussian distribution and exhibit a multimodal distribution, which violates the diffusion assumption and degrades the results. To overcome this limitation, we propose a fast SRDiff (FSRDiff) algorithm that integrates a generative adversarial network (GAN) with a diffusion model to speed up SRDiff. FSRDiff employs conditional GAN to approximate the multimodal distribution in the inverse diffusion process of the diffusion model, thus enhancing its sampling efficiency when reducing the total number of diffusion steps. The experimental results show that FSRDiff is nearly 20 times faster than SRDiff in reconstruction while maintaining comparable performance on the DIV2K test set.

Improved PAA algorithm for breast mass detection in mammograms

Mammography screening is instrumental in the early detection and diagnosis of breast cancer by identifying masses in mammograms. With the rapid development of deep learning, numerous deep learning-based object detection algorithms have been explored for mass detection studies. However, these methods often yield a high false positive rate per image (FPPI) while achieving a high true positive rate (TPR). To maintain a higher TPR while also ensuring lower FPPI, we improved the Probability Anchor Assignment (PAA) algorithm to enhance the detection capability for mammographic characteristics with our previous work. We considered three dimensions: the backbone network, feature fusion module, and dense detection heads. The final experiment showed the effectiveness of the proposed method, and the TPR/FPPI values of the final improved PAA algorithm were 0.96/0.56 on the INbreast datasets. Compared to other methods, our method stands distinguished with its effectiveness in addressing the imbalance between positive and negative classes in cases of single lesion detection.

CSDFormer: A cloud and shadow detection method for landsat images based on transformer

Cloud and shadow (CS) detection is crucial prerequisite for application of remote sensing images. Current deep learning-based detection algorithms mainly employ Convolutional Neural Networks (CNNs). However, the local receptive field in CNNs cannot effectively capture global contextual information, which hinders accurate characterization of the dependency between clouds and shadows. In vision Transformers, self-attention mechanisms can effectively capture the long-distance dependencies between different regions in an image. Inspired by this, this paper proposed a new CS Detection algorithm based on a Transformer, called CSDFormer. Specifically, we exclusively employed a hierarchical Transformer structure in the encoder stage to extract features of CS. Each Transformer layer contains several multi-head self-attention mechanisms for calculating pixel-wise long-distance connectivity. The designed structure enables the Transformer to better extract global context information, which helps to strengthen the comprehension of the semantic relationships between clouds and shadows. Benefiting from the global feature extraction capability of the encoder stage, we employed several simple multilayer perceptron layers for multi-scale feature map fusion and pixel classification in the decoder stage. The proposed CSDFormer was validated using 898 Landsat 8 Biome images with 512 × 512 pixels, producing an overall accuracy of 95.28 % and a mean intersection over union of 84.08 %, outperforming three state-of-the-art CNN-based algorithms. CSDFormer is consistently more accurate in detection of both clouds and shadows. Owing to the parallel computing capability of the self-attention mechanism, CSDFormer is computationally more efficient than the three CNN-based benchmark methods. For the input spectral bands, the performance of CSDFormer produced can be further enhanced with additional thermal infrared bands.

YOLO-Ships: Lightweight ship object detection based on feature enhancement

Due to the influences of complex shore-based backgrounds, ship scale changes and occlusions, most deep learning-based ship target detection algorithms yield false and missed detections, and the number of required model parameters is large. Therefore, we present a lightweight and effective method called YOLO-Ships. An improved lightweight Ghost module was used to enhance the ability of the network to extract ship features. In addition, we designed a ship feature enhancement module and a lightweight C3REGhost module to strengthen the important feature information and improve the detection ability and localization accuracy of the model. Finally, the model loss function was redesigned to improve the convergence speed and accuracy of the model. Our experimental results showed that the mAP@0.5 increased by 5.2% and the number of parameters decreased by 2.89 M compared to those of YOLOv5s, effectively achieving a balance between model lightweightness and good ship detection accuracy.

Evaluation of a Cascaded Deep Learning-based Algorithm for Prostate Lesion Detection at Biparametric MRI.

Background Multiparametric MRI (mpMRI) improves prostate cancer (PCa) detection compared with systematic biopsy, but its interpretation is prone to interreader variation, which results in performance inconsistency. Artificial intelligence (AI) models can assist in mpMRI interpretation, but large training data sets and extensive model testing are required. Purpose To evaluate a biparametric MRI AI algorithm for intraprostatic lesion detection and segmentation and to compare its performance with radiologist readings and biopsy results. Materials and Methods This secondary analysis of a prospective registry included consecutive patients with suspected or known PCa who underwent mpMRI, US-guided systematic biopsy, or combined systematic and MRI/US fusion-guided biopsy between April 2019 and September 2022. All lesions were prospectively evaluated using Prostate Imaging Reporting and Data System version 2.1. The lesion- and participant-level performance of a previously developed cascaded deep learning algorithm was compared with histopathologic outcomes and radiologist readings using sensitivity, positive predictive value (PPV), and Dice similarity coefficient (DSC). Results A total of 658 male participants (median age, 67 years [IQR, 61-71 years]) with 1029 MRI-visible lesions were included. At histopathologic analysis, 45% (294 of 658) of participants had lesions of International Society of Urological Pathology (ISUP) grade group (GG) 2 or higher. The algorithm identified 96% (282 of 294; 95% CI: 94%, 98%) of all participants with clinically significant PCa, whereas the radiologist identified 98% (287 of 294; 95% CI: 96%, 99%; P = .23). The algorithm identified 84% (103 of 122), 96% (152 of 159), 96% (47 of 49), 95% (38 of 40), and 98% (45 of 46) of participants with ISUP GG 1, 2, 3, 4, and 5 lesions, respectively. In the lesion-level analysis using radiologist ground truth, the detection sensitivity was 55% (569 of 1029; 95% CI: 52%, 58%), and the PPV was 57% (535 of 934; 95% CI: 54%, 61%). The mean number of false-positive lesions per participant was 0.61 (range, 0-3). The lesion segmentation DSC was 0.29. Conclusion The AI algorithm detected cancer-suspicious lesions on biparametric MRI scans with a performance comparable to that of an experienced radiologist. Moreover, the algorithm reliably predicted clinically significant lesions at histopathologic examination. ClinicalTrials.gov Identifier: NCT03354416 © RSNA, 2024 Supplemental material is available for this article.

Advanced deep learning-based image reconstruction in lumbar spine MRI at 0.55 T – Effects on image quality and acquisition time in comparison to conventional deep learning-based reconstruction

ObjectivesTo evaluate an optimized deep leaning-based image post-processing technique in lumbar spine MRI at 0.55 T in terms of image quality and image acquisition time. Materials and methodsLumbar spine imaging was conducted on 18 patients using a 0.55 T MRI scanner, employing conventional (CDLR) and advanced (ADLR) deep learning-based post-processing techniques. Two musculoskeletal radiologists visually evaluated the images using a 5-point Likert scale to assess image quality and resolution. Quantitative assessment in terms of signal intensities (SI) and contrast ratios was performed by region of interest measurements in different body-tissues (vertebral bone, intervertebral disc, spinal cord, cerebrospinal fluid and autochthonous back muscles) to investigate differences between CDLR and ADLR sequences. ResultsThe images processed with the advanced technique (ADLR) were rated superior to the conventional technique (CDLR) in terms of signal/contrast, resolution, and assessability of the spinal canal and neural foramen. The interrater agreement was moderate for signal/contrast (ICC = 0.68) and good for resolution (ICC = 0.77), but moderate for spinal canal and neuroforaminal assessability (ICC = 0.55). Quantitative assessment showed a higher contrast ratio for fluid-sensitive sequences in the ADLR images. The use of ADLR reduced image acquisition time by 44.4%, from 14:22 min to 07:59 min. ConclusionsAdvanced deep learning-based image reconstruction algorithms improve the visually perceived image quality in lumbar spine imaging at 0.55 T while simultaneously allowing to substantially decrease image acquisition times. Clinical relevanceAdvanced deep learning-based image post-processing techniques (ADLR) in lumbar spine MRI at 0.55 T significantly improves image quality while reducing image acquisition time.

Deep Reinforcement Learning based channel allocation (DRLCA) in Cognitive Radio Networks

The Spectrum is a scarce resource in wireless networks, hence these resources need to be perfectly channelized for their better utilization. By the study of the previous decades, it is found that the spectrum is colossally underutilized and the main reason for the underutilization found out to be the policies that are fixed and not dynamic. The dynamic spectrum allocation of frequency bands may overcome this problem. Cognitive radio provides an important concept that can used to solve the problem of underutilization of spectrum. Reinforcement learning is a key technique that is widely used to learn the spectrum allocation behaviour and maximize the system’s efficiency. Therefore, in this work, we have designed and developed a Reinforcement learning-based model to allocate the channels among secondary users. Also, a Deep Reinforcement learning-based channel allocation algorithm (DRLCA) has been proposed. The proposed DRLCA is compared with existing JPCRL [1]. In our algorithm, the Python libraries were used for simulation. From the simulation results and analysis, it is found that the DRLCA outperforms the JPCRL in terms of channel utilization by 5%.

Real-Time Multi-Person Video Synthesis with Controllable Prior-Guided Matting.

In order to enhance the matting performance in multi-person dynamic scenarios, we introduce a robust, real-time, high-resolution, and controllable human video matting method that achieves state of the art on all metrics. Unlike most existing methods that perform video matting frame by frame as independent images, we design a unified architecture using a controllable generation model to solve the problem of the lack of overall semantic information in multi-person video. Our method, called ControlMatting, uses an independent recurrent architecture to exploit temporal information in videos and achieves significant improvements in temporal coherence and detailed matting quality. ControlMatting adopts a mixed training strategy comprised of matting and a semantic segmentation dataset, which effectively improves the semantic understanding ability of the model. Furthermore, we propose a novel deep learning-based image filter algorithm that enforces our detailed augmentation ability on both matting and segmentation objectives. Our experiments have proved that prior information about the human body from the image itself can effectively combat the defect masking problem caused by complex dynamic scenarios with multiple people.

On the design of deep learning-based control algorithms for visually guided UAVs engaged in power tower inspection tasks.

This paper focuses on the design of Convolution Neural Networks to visually guide an autonomous Unmanned Aerial Vehicle required to inspect power towers. The network is required to precisely segment images taken by a camera mounted on a UAV in order to allow a motion module to generate collision-free and inspection-relevant manoeuvres of the UAV along different types of towers. The images segmentation process is particularly challenging not only because of the different structures of the towers but also because of the enormous variability of the background, which can vary from the uniform blue of the sky to the multi-colour complexity of a rural, forest, or urban area. To be able to train networks that are robust enough to deal with the task variability, without incurring into a labour-intensive and costly annotation process of physical-world images, we have carried out a comparative study in which we evaluate the performances of networks trained either with synthetic images (i.e., the synthetic dataset), physical-world images (i.e., the physical-world dataset), or a combination of these two types of images (i.e., the hybrid dataset). The network used is an attention-based U-NET. The synthetic images are created using photogrammetry, to accurately model power towers, and simulated environments modelling a UAV during inspection of different power towers in different settings. Our findings reveal that the network trained on the hybrid dataset outperforms the networks trained with the synthetic and the physical-world image datasets. Most notably, the networks trained with the hybrid dataset demonstrates a superior performance on multiples evaluation metrics related to the image-segmentation task. This suggests that, the combination of synthetic and physical-world images represents the best trade-off to minimise the costs related to capturing and annotating physical-world images, and to maximise the task performances. Moreover, the results of our study demonstrate the potential of photogrammetry in creating effective training datasets to design networks to automate the precise movement of visually-guided UAVs.

A Systematic Review of Electroencephalography-Based Emotion Recognition of Confusion Using Artificial Intelligence

Confusion emotion in a learning environment can motivate the learner, but prolonged confusion hinders the learning process. Recognizing confused learners is possible; nevertheless, finding them requires a lot of time and effort. Due to certain restrictions imposed by the settings of an online learning environment, the recognition of confused students is a big challenge for educators. Therefore, novel technologies are necessary to handle such crucial difficulties. Lately, Electroencephalography (EEG)-based emotion recognition systems have been rising in popularity in the domain of Education Technology. Such systems have been utilized to recognize the confusion emotion of learners. Numerous studies have been conducted to recognize confusion emotion through this system since 2013, and because of this, a systematic review of the methodologies, feature sets, and utilized classifiers is a timely necessity. This article presents the findings of the review conducted to achieve this requirement. We summarized the published literature in terms of the utilized datasets, feature preprocessing, feature types for model training, and deployed classifiers in terms of shallow machine learning and deep learning-based algorithms. Moreover, the article presents a comparison of the prediction accuracies of the classifiers and illustrates the existing research gaps in confusion emotion recognition systems. Future study directions for potential research are also suggested to overcome existing gaps.

Deep Learning for Cell Migration in Nonwoven Materials and Evaluating Gene Transfer Effects following AAV6-ND4 Transduction.

Studying cell settlement in the three-dimensional structure of synthetic biomaterials over time is of great interest in research and clinical translation for the development of artificial tissues and organs. Tracking cells as physical objects improves our understanding of the processes of migration, homing, and cell division during colonisation of the artificial environment. In this study, the 3D environment had a direct effect on the behaviour of biological objects. Recently, deep learning-based algorithms have shown significant benefits for cell segmentation tasks and, furthermore, for biomaterial design optimisation. We analysed the primary LHON fibroblasts in an artificial 3D environment after adeno-associated virus transduction. Application of these tools to model cell homing in biomaterials and to monitor cell morphology, migration and proliferation indirectly demonstrated restoration of the normal cell phenotype after gene manipulation by AAV transduction. Following the 3Rs principles of reducing the use of living organisms in research, modeling the formation of tissues and organs by reconstructing the behaviour of different cell types on artificial materials facilitates drug testing, the study of inherited and inflammatory diseases, and wound healing. These studies on the composition and algorithms for creating biomaterials to model the formation of cell layers were inspired by the principles of biomimicry.

Enhanced bone assessment of the shoulder using zero-echo time MRI with deep-learning image reconstruction.

To assess a deep learning-based reconstruction algorithm (DLRecon) in zero echo-time (ZTE) MRI of the shoulder at 1.5 Tesla for improved delineation of osseous findings. In this retrospective study, 63 consecutive exams of 52 patients (28 female) undergoing shoulder MRI at 1.5 Tesla in clinical routine were included. Coronal 3D isotropic radial ZTE pulse sequences were acquired in the standard MR shoulder protocol. In addition to standard-of-care (SOC) image reconstruction, the same raw data was reconstructed with a vendor-supplied prototype DLRecon algorithm. Exams were classified into three subgroups: no pathological findings, degenerative changes, and posttraumatic changes, respectively. Two blinded readers performed bone assessment on a 4-point scale (0-poor, 3-perfect) by qualitatively grading image quality features and delineation of osseous pathologies including diagnostic confidence in the respective subgroups. Quantitatively, signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of bone were measured. Qualitative variables were compared using the Wilcoxon signed-rank test for ordinal data and the McNemar test for dichotomous variables; quantitative measures were compared with Student's t-testing. DLRecon scored significantly higher than SOC in all visual metrics of image quality (all, p < 0.03), except in the artifact category (p = 0.37). DLRecon also received superior qualitative scores for delineation of osseous pathologies and diagnostic confidence (p ≤ 0.03). Quantitatively, DLRecon achieved superior CNR (95 CI [1.4-3.1]) and SNR (95 CI [15.3-21.5]) of bone than SOC (p < 0.001). DLRecon enhanced image quality in ZTE MRI and improved delineation of osseous pathologies, allowing for increased diagnostic confidence in bone assessment.

The Detection of Pulp Stones with Automatic Deep Learning in Panoramic Radiographies: An AI Pilot Study.

This study aims to evaluate the effectiveness of employing a deep learning approach for the automated detection of pulp stones in panoramic imaging. A comprehensive dataset comprising 2409 panoramic radiography images (7564 labels) underwent labeling using the CranioCatch labeling program, developed in Eskişehir, Turkey. The dataset was stratified into three distinct subsets: training (n = 1929, 80% of the total), validation (n = 240, 10% of the total), and test (n = 240, 10% of the total) sets. To optimize the visual clarity of labeled regions, a 3 × 3 clash operation was applied to the images. The YOLOv5 architecture was employed for artificial intelligence modeling, yielding F1, sensitivity, and precision metrics of 0.7892, 0.8026, and 0.7762, respectively, during the evaluation of the test dataset. Among deep learning-based artificial intelligence algorithms applied to panoramic radiographs, the use of numerical identification for the detection of pulp stones has achieved remarkable success. It is expected that the success rates of training models will increase by using datasets consisting of a larger number of images. The use of artificial intelligence-supported clinical decision support system software has the potential to increase the efficiency and effectiveness of dentists.

Multi-task learning for IoT traffic classification: A comparative analysis of deep autoencoders

As a system allowing intra-network devices to automatically communicate over the Internet, the Internet of Things (IoT) faces increasing popularity in modern applications and security threats — particularly network intrusions that target both networks and devices. A major threat is network attacks that attempt to obtain unauthorised access and damage the networks or systems. To effectively safeguard against these attacks, it is essential to use efficient hybrid intrusion detection systems with advanced techniques. Deep learning-based algorithms, such as Autoencoders (AEs), have been recognised as efficient methods for intrusion detection. However, it is crucial to further analyse the impact of different structures on detection performance. This paper proposes a method combining AE for data reconstruction and anomaly detection and multi-task learning (MTL)-structured models for IoT traffic classification. Additionally, we suggest a hybrid resampling technique with the Synthetic Minority Over-Sampling Technique and Generative Adversarial Network (SMOTE-GAN) for enhanced training and conduct a comparative analysis of different neural networks with AEs. The experimental results on two publicly available datasets demonstrate the effectiveness and practicality of the proposed AE-MTL approach compared with single-task learning. Additionally, while the performance varies on different datasets, Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU), and Convolutional Neural Network (CNN) have improved the detection of minor classes.

Automatic Defect Detection of Jet Engine Turbine and Compressor Blade Surface Coatings Using a Deep Learning-Based Algorithm

Automatic Defect Detection of Jet Engine Turbine and Compressor Blade Surface Coatings Using a Deep Learning-Based Algorithm

Fault transfer diagnosis of rolling bearings across different devices via multi-domain information fusion and multi-kernel maximum mean discrepancy

The current deep learning-based intelligent diagnosis algorithms depend on large amounts of well-labeled data, but they may not perform well in engineering practice where the fault data available from a single device is limited. The significant difference in data distribution between different devices makes it more difficult to transfer diagnosis across devices. Therefore, aiming at the problem that the great difference in the data distribution of source and target domains in cross-device transfer diagnosis leads to the low diagnosis accuracy, this paper investigates a deep transfer network model based on multi-domain information fusion and multi-kernel maximum mean discrepancy (MK-MMD) for fault diagnosis of rolling bearings across different devices. Firstly, to address the problem that single-domain information is difficult to adequately characterize rolling bearing health status information of different devices, a multi-domain information feature extractor based on multi-attention mechanism is constructed to capture the important features of vibration signals in different transform domains. Secondly, the extracted multi-domain features are input into the bidirectional gated recurrent unit (BiGRU) network for feature fusion, and MK-MMD is used to narrow the distribution distance between the source domain and target domain to obtain domain invariant features, and then achieve the transfer diagnosis of rolling bearing faults between different devices. Finally, the method investigated is verified by rolling bearing fault transfer diagnosis tests between different devices, and the results show that the suggested method can adapt to the feature distribution between different domains, and improve the transfer diagnosis accuracy between different devices.