Improving artificial intelligence pipeline for liver malignancy diagnosis using ultrasound images and video frames.

Recent developments of deep learning methods have demonstrated their feasibility in liver malignancy diagnosis using ultrasound (US) images. However, most of these methods require manual selection and annotation of US images by radiologists, which limit their practical application. On the other hand, US videos provide more comprehensive morphological information about liver masses and their relationships with surrounding structures than US images, potentially leading to a more accurate diagnosis. Here, we developed a fully automated artificial intelligence (AI) pipeline to imitate the workflow of radiologists for detecting liver masses and diagnosing liver malignancy. In this pipeline, we designed an automated mass-guided strategy that used segmentation information to direct diagnostic models to focus on liver masses, thus increasing diagnostic accuracy. The diagnostic models based on US videos utilized bi-directional convolutional long short-term memory modules with an attention-boosted module to learn and fuse spatiotemporal information from consecutive video frames. Using a large-scale dataset of 50 063 US images and video frames from 11 468 patients, we developed and tested the AI pipeline and investigated its applications. A dataset of annotated US images is available at https://doi.org/10.5281/zenodo.7272660.

Molecular mechanism of brassinosteroids involved in root gravity response based on transcriptome analysis.

BACKGROUND: Brassinosteroids (BRs) are a class of phytohormones that regulate a wide range of developmental processes in plants. BR-associated mutants display impaired growth and response to developmental and environmental stimuli. RESULTS: Here, we found that a BR-deficient mutant det2-1 displayed abnormal root gravitropic growth in Arabidopsis, which was not present in other BR mutants. To further elucidate the role of DET2 in gravity, we performed transcriptome sequencing and analysis of det2-1 and bri1-116, bri1 null mutant allele. Expression levels of auxin, gibberellin, cytokinin, and other related genes in the two mutants of det2-1 and bri1-116 were basically the same. However, we only found that a large number of JAZ (JASMONATE ZIM-domain) genes and jasmonate synthesis-related genes were upregulated in det2-1 mutant, suggesting increased levels of endogenous JA. CONCLUSIONS: Our results also suggested that DET2 not only plays a role in BR synthesis but may also be involved in JA regulation. Our study provides a new insight into the molecular mechanism of BRs on the root gravitropism.

Kinase regulators evolved into two families by gain and loss of ability to bind plant steroid receptors.

All biological functions evolve by fixing beneficial mutations and removing deleterious ones. Therefore, continuously fixing and removing the same essential function to separately diverge monophyletic gene families sounds improbable. Yet, here we report that brassinosteroid insensitive1 kinase inhibitor1 (BKI1)/membrane-associated kinase regulators (MAKRs) regulating a diverse function evolved into BKI1 and MAKR families from a common ancestor by respectively enhancing and losing ability to bind brassinosteroid receptor brassinosteroid insensitive1 (BRI1). The BKI1 family includes BKI1, MAKR1/BKI1-like (BKL) 1, and BKL2, while the MAKR family contains MAKR2-6. Seedless plants contain only BKL2. In seed plants, MAKR1/BKL1 and MAKR3, duplicates of BKL2, gained and lost the ability to bind BRI1, respectively. In angiosperms, BKL2 lost the ability to bind BRI1 to generate MAKR2, while BKI1 and MAKR6 were duplicates of MAKR1/BKL1 and MAKR3, respectively. In dicots, MAKR4 and MAKR5 were duplicates of MAKR3 and MAKR2, respectively. Importantly, BKI1 localized in the plasma membrane, but BKL2 localized to the nuclei while MAKR1/BKL1 localized throughout the whole cell. Importantly, BKI1 strongly and MAKR1/BKL1 weakly inhibited plant growth, but BKL2 and the MAKR family did not inhibit plant growth. Functional study of the chimeras of their N- and C-termini showed that only the BKI1 family was partially reconstructable, supporting stepwise evolution by a seesaw mechanism between their C- and N-termini to alternately gain an ability to bind and inhibit BRI1, respectively. Nevertheless, the C-terminal BRI1-interacting motif best defines the divergence of BKI1/MAKRs. Therefore, BKI1 and MAKR families evolved by gradually gaining and losing the same function, respectively, extremizing divergent evolution and adding insights into gene (BKI1/MAKR) duplication and divergence.

Eigendecomposition-free inverse design of meta-optics devices.

The inverse design of meta-optics has received much attention in recent years. In this paper, we propose a GPU-friendly inverse design framework based on improved eigendecomposition-free rigorous diffraction interface theory, which offers up to 16.2 × speedup over the traditional inverse design based on rigorous coupled-wave analysis. We further improve the framework's flexibility by introducing a hybrid parameterization combining neural-implicit and traditional shape optimization. We demonstrate the effectiveness of our framework through intricate tasks, including the inverse design of reconfigurable free-form meta-atoms.

An ultrasound-based ensemble machine learning model for the preoperative classification of pleomorphic adenoma and Warthin tumor in the parotid gland.

OBJECTIVES: The preoperative classification of pleomorphic adenomas (PMA) and Warthin tumors (WT) in the parotid gland plays an essential role in determining therapeutic strategies. This study aims to develop and validate an ultrasound-based ensemble machine learning (USEML) model, employing nonradiative and noninvasive features to differentiate PMA from WT. METHODS: A total of 203 patients with histologically confirmed PMA or WT who underwent parotidectomy from two centers were enrolled. Clinical factors, ultrasound (US) features, and radiomic features were extracted to develop three types of machine learning model: clinical models, US models, and USEML models. The diagnostic performance of the USEML model, as well as that of physicians based on experience, was evaluated and validated using receiver operating characteristic (ROC) curves in internal and external validation cohorts. DeLong's test was used for comparisons of AUCs. SHAP values were also utilized to explain the classification model. RESULTS: The USEML model achieved the highest AUC of 0.891 (95% CI, 0.774-0.961), surpassing the AUCs of both the US (0.847; 95% CI, 0.720-0.932) and clinical (0.814; 95% CI, 0.682-0.908) models. The USEML model also outperformed physicians in both internal and external validation datasets (both p < 0.05). The sensitivity, specificity, negative predictive value, and positive predictive value of the USEML model and physician experience were 89.3%/75.0%, 87.5%/54.2%, 87.5%/65.6%, and 89.3%/65.0%, respectively. CONCLUSIONS: The USEML model, incorporating clinical factors, ultrasound factors, and radiomic features, demonstrated efficient performance in distinguishing PMA from WT in the parotid gland. CLINICAL RELEVANCE STATEMENT: This study developed a machine learning model for preoperative diagnosis of pleomorphic adenoma and Warthin tumor in the parotid gland based on clinical, ultrasound, and radiomic features. Furthermore, it outperformed physicians in an external validation dataset, indicating its potential for clinical application. KEY POINTS: • Differentiating pleomorphic adenoma (PMA) and Warthin tumor (WT) affects management decisions and is currently done by invasive biopsy. • Integration of US-radiomic, clinical, and ultrasound findings in a machine learning model results in improved diagnostic accuracy. • The ultrasound-based ensemble machine learning (USEML) model consistently outperforms physicians, suggesting its potential applicability in clinical settings.

Human periodontal ligament stem cell sheets activated by graphene oxide quantum dots repair periodontal bone defects by promoting mitochondrial dynamics dependent osteogenic differentiation.

BACKGROUND: Bone defects in the maxillofacial region restrict the integrity of dental function, posing challenges in clinical treatment. Bone tissue engineering (BTE) with stem cell implants is an effective method. Nanobiomaterials can effectively enhance the resistance of implanted stem cells to the harsh microenvironment of bone defect areas by promoting cell differentiation. Graphene oxide quantum dots (GOQDs) are zero-dimensional nanoscale derivatives of graphene oxide with excellent biological activity. In the present study, we aimed to explore the effects of GOQDs prepared by two methods (Y-GOQDs and B-GOQDs) on the osteogenic differentiation of human periodontal ligament stem cells (hPDLSCs), as well as the effect of gelatin methacryloyl (GelMA)-encapsulated GOQD-induced hPDLSC sheets on the repair of mandibular periodontal defects in rats. We also explored the molecular biological mechanism through which GOQD promotes bone differentiation. RESULTS: There were significant differences in oxygen-containing functional groups, particle size and morphology between Y-GOQDs and B-GOQDs. Y-GOQDs promoted the osteogenic differentiation of hPDLSCs more effectively than did B-GOQDs. In addition, GelMA hydrogel-encapsulated Y-GOQD-induced hPDLSC cell sheet fragments not only exhibited good growth and osteogenic differentiation in vitro but also promoted the repair of mandibular periodontal bone defects in vivo. Furthermore, the greater effectiveness of Y-GOQDs than B-GOQDs in promoting osteogenic differentiation is due to the regulation of hPDLSC mitochondrial dynamics, namely, the promotion of fusion and inhibition of fission. CONCLUSIONS: Overall, Y-GOQDs are more effective than B-GOQDs at promoting the osteogenic differentiation of hPDLSCs by regulating mitochondrial dynamics, which ultimately contributes to bone regeneration via the aid of the GelMA hydrogels in vivo.

Automatic segmentation of 15 critical anatomical labels and measurements of cardiac axis and cardiothoracic ratio in fetal four chambers using nnU-NetV2.

BACKGROUND: Accurate segmentation of critical anatomical structures in fetal four-chamber view images is essential for the early detection of congenital heart defects. Current prenatal screening methods rely on manual measurements, which are time-consuming and prone to inter-observer variability. This study develops an AI-based model using the state-of-the-art nnU-NetV2 architecture for automatic segmentation and measurement of key anatomical structures in fetal four-chamber view images. METHODS: A dataset, consisting of 1,083 high-quality fetal four-chamber view images, was annotated with 15 critical anatomical labels and divided into training/validation (867 images) and test (216 images) sets. An AI-based model using the nnU-NetV2 architecture was trained on the annotated images and evaluated using the mean Dice coefficient (mDice) and mean intersection over union (mIoU) metrics. The model's performance in automatically computing the cardiac axis (CAx) and cardiothoracic ratio (CTR) was compared with measurements from sonographers with varying levels of experience. RESULTS: The AI-based model achieved a mDice coefficient of 87.11% and an mIoU of 77.68% for the segmentation of critical anatomical structures. The model's automated CAx and CTR measurements showed strong agreement with those of experienced sonographers, with respective intraclass correlation coefficients (ICCs) of 0.83 and 0.81. Bland-Altman analysis further confirmed the high agreement between the model and experienced sonographers. CONCLUSION: We developed an AI-based model using the nnU-NetV2 architecture for accurate segmentation and automated measurement of critical anatomical structures in fetal four-chamber view images. Our model demonstrated high segmentation accuracy and strong agreement with experienced sonographers in computing clinically relevant parameters. This approach has the potential to improve the efficiency and reliability of prenatal cardiac screening, ultimately contributing to the early detection of congenital heart defects.

LINC01572 promotes triple-negative breast cancer progression through EIF4A3-mediated ß-catenin mRNA nuclear exportation.

Long noncoding RNAs have been reported to be involved in the development of breast cancer. LINC01572 was previously reported to promote the development of various tumors. However, the potential biological function of LINC01572 in breast cancer remains largely unknown. R language was used to perform bioinformatic analysis of The Cancer Genome Atlas data. The expression level of RNAs was examined by RT-qPCR. The effect of knocking down or overexpression LINC01572 in triple-negative breast cancer (TNBC) cell lines was evaluated by detecting cell proliferation, migrant action. RNA immunoprecipitation assay and RNA pull-down assay were performed to explore the regulatory relationship between LINC01572, EIF4A3, and ß-catenin. Bioinformatics analysis identifies LINC01572 as an oncogene of breast cancer. LINC01572 is over-expressed in TNBC tissues and cell lines, correlated with poor clinical prognosis in BC patients. Cell function studies confirmed that LINC01572 facilitated the proliferation and migration of TNBC cells in both vivo and vitro. Mechanistically, ß-catenin mRNA and EIF4A3 combine spatially to form a complex, LINC01572 helps transport this complex from the nucleus to the cytoplasm, thereby facilitating the translation of ß-catenin. Our findings confirm that LINC01572 acts as a tumor promoter and may act as a biomarker in TNBC. In addition, novel molecular regulatory relationships involving LINC01572/EIF4A3/ß-catenin are critical to the development of TNBC, which led to a new understanding of the mechanisms of TNBC progression and shows a new target for precision treatment for TNBC.

GBDT Method Integrating Feature-Enhancement and Active-Learning Strategies-Sea Ice Thickness Inversion in Beaufort Sea.

Sea ice, as an important component of the Earth's ecosystem, has a profound impact on global climate and human activities due to its thickness. Therefore, the inversion of sea ice thickness has important research significance. Due to environmental and equipment-related limitations, the number of samples available for remote sensing inversion is currently insufficient. At high spatial resolutions, remote sensing data contain limited information and noise interference, which seriously affect the accuracy of sea ice thickness inversion. In response to the above issues, we conducted experiments using ice draft data from the Beaufort Sea and designed an improved GBDT method that integrates feature-enhancement and active-learning strategies (IFEAL-GBDT). In this method, the incident angle and time series are used to perform spatiotemporal correction of the data, reducing both temporal and spatial impacts. Meanwhile, based on the original polarization information, effective multi-attribute features are generated to expand the information content and improve the separability of sea ice with different thicknesses. Taking into account the growth cycle and age of sea ice, attributes were added for month and seawater temperature. In addition, we studied an active learning strategy based on the maximum standard deviation to select more informative and representative samples and improve the model's generalization ability. The improved GBDT model was used for training and prediction, offering advantages in dealing with nonlinear, high-dimensional data, and data noise problems, further expanding the effectiveness of feature-enhancement and active-learning strategies. Compared with other methods, the method proposed in this paper achieves the best inversion accuracy, with an average absolute error of 8 cm and a root mean square error of 13.7 cm for IFEAL-GBDT and a correlation coefficient of 0.912. This research proves the effectiveness of our method, which is suitable for the high-precision inversion of sea ice thickness determined using Sentinel-1 data.

Three-dimensional finite element analysis of the uprighting movement of mandibular mesially inclined second molars.

INTRODUCTION: The mesially inclined mandibular second molar can be supported upright by the microimplant anchorage. This study established the finite element model to analyze the displacement trend and periodontal ligament (PDL) stress distribution of the uprighting mandibular second molar with the microimplant under different conditions. METHODS: A 3-dimensional model of the mandible and dentition was established. The mesial inclination of the mandibular second molar was 30°, 45°, and 60°. Microimplants were implanted between the buccal side of the second premolar and the first molar and in the distal part of the mandibular second molar, respectively. Six groups were set, each loaded with 0.5 N of force. The second molar initial displacement trend and PDL stress distribution were evaluated. RESULTS: The PDL stress of mandibular second molars in all groups was within the physiological limit, and the PDL stress of mandibular second molars in the distal implant groups was lower than that of mandibular second molars in the mesial implant groups. PDL stress concentration in the cervical area. Tooth displacement decreased as the mesial inclination angle of the second molars increased. The sagittal displacement of mesial implant groups was larger, and there was a tendency of mesiobuccal torsion when standing uprighting; the vertical displacement of distal implant groups was larger, and there was a tendency of distal lingual torsion when standing uprighting. CONCLUSIONS: Distal microimplant has a better extrusion effect on the mesially inclined second molar, whereas mesial microimplant has a better effect on the distal movement. The optimal orthodontic force for microimplant traction on mesially inclined second molars is 0.5-0.8 N.

Impact of molar teeth distalization by clear aligners on maxillary alveolar bone thickness and root resorption: a three­dimensional study.

OBJECTIVE: This study aimed to evaluate the impact of molar teeth distalization movement by clear aligners on changes in the alveolar bone thickness and orthodontically induced inflammatory root resorption (OIIRR) in maxillary molars using conebeam computed tomography (CBCT). MATERIALS AND METHODS: Three-dimensional CBCT scans of 35 adult patients (one hundred forty maxillary molars) with pre-designed selection criteria and a mean age of 24.4 ± 7.1 years were included. The measured parameters, including alveolar bone thickness for maxillary molars and root resorption (OIIRR), were analyzed using pre-and post-treatment CBCT (T0 and T1, respectively) with Invivo 6.0 software. RESULT: Post-treatment, relevant statistically significant changes included deposition of bone in the average palatal surface of the 1st molars. The reduction of bone was seen in the average buccal surface of the first molars and both surfaces of the second molars. Regarding root length after treatment, the average maxillary 1st molar roots showed significant OIIRR (p < 0.001). CONCLUSION: Clear aligner treatment could effectively reduce the incidence of alveolar bone thickness reduction and OIIRR in treating Class II malocclusions compared to conventional braces, as shown in previous studies. This research will aid in fully grasping the benefits of clear aligners.

Novel Strategy to Rapidly Profile and Identify Oxidized Species of Polysorbate 80 Using Ultra-High-Performance Liquid Chromatography Coupled with High-Resolution Mass Spectrometry.

Polysorbate 80 (PS80) is widely used as an excipient in vaccines and biopharmaceuticals. The oxidized species of PS80 have raised concern because of their potential to compromise product stability and pose a clinical risk. Analytical methods to profile and identify the oxidized species are hard to develop owing to their complexity and low abundance. Herein, a novel strategy was demonstrated to comprehensively profile and identify the oxidized species of PS80 using ultra-high-performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry. The characteristic fragmentation patterns of the oxidized species were obtained under the "all ions" scan mode. Then, 10 types of distinct fragments from oxidized oleates were identified and confirmed using two purified oxidized species (polyoxyethylene (POE) sorbitan mono-hydroxy oleate and POE mono-keto oleate) whose structures were elucidated via nuclear magnetic resonance. A total of 348 (32 types) oxidized species were profiled and identified in the oxidized PS80 samples, including 119 (10 types) species found for the first time to our knowledge. Mathematical models were established and validated based on the good logarithmic relation between the POE degree of polymerization and the relative retention time and used to rapidly discover and identify the oxidized species. A novel strategy was established to profile and identify the PS80 oxidized species based on their retention time, HRMS, and HRMS2 data of the detected peaks using an in-house dataset. Using this strategy, 104 (14 types) and 97 (13 types) oxidized species were identified for the first time in PS80 and its preparations, respectively.

Expression and regulatory network of E3 ubiquitin ligase NEDD4 family in cancers.

NEDD4 family represent an important group of E3 ligases, which regulate various cellular pathways of cell proliferation, cell junction and inflammation. Emerging evidence suggested that NEDD4 family members participate in the initiation and development of tumor. In this study, we systematically investigated the molecular alterations as well as the clinical relevance regarding NEDD4 family genes in 33 cancer types. Finally, we found that NEDD4 members showed increased expression in pancreas cancer and decreased expression in thyroid cancer. NEDD4 E3 ligase family genes had an average mutation frequency in the range of 0-32.1%, of which HECW1 and HECW2 demonstrated relatively high mutation rate. Breast cancer harbors large amount of NEDD4 copy number amplification. NEDD4 family members interacted proteins were enriched in various pathways including p53, Akt, apoptosis and autophagy, which were confirmed by further western blot and flow cytometric analysis in A549 and H1299 lung cancer cells. In addition, expression of NEDD4 family genes were associated with survival of cancer patients. Our findings provide novel insight into the effect of NEDD4 E3 ligase genes on cancer progression and treatment in the future.

Uniform Two-Dimensional Crystalline Platelets with Tailored Compositions for pH Stimulus-Responsive Drug Release.

Two-dimensional, size-tunable, water-dispersible particle micelles with spatially defined chemistries can be obtained by using "living" crystallization-driven self-assembly (CDSA) approach. Nevertheless, a major obstacle of crystalline particles in drug delivery application is the difficulty in accessing to cargo within crystalline cores. In the present work, we design four different types of biocompatible two-dimensional platelets with a crystalline poly(ε-caprolactone) (PCL) core, a hydrophobic poly(4-vinylprydine) (P4VP) segment, and a water dispersible poly(N,N-dimethyl acrylamide) (PDMA) block in ethanol by seeded growth method. Transferring those uniform platelets with tailored compositions to an aqueous solution in the presence of a hydrophobic drug leads to efficient encapsulation of the cargo in the P4VP segments via hydrophobic interactions. These drug-loaded platelets exhibit pH-responsive release behavior in aqueous media due to the protonated-deprotonated process of P4VP blocks in acidic and neutral solutions. This work provides initial insight into biocompatible PCL platelets with low dispersity and precise chemistry control in stimulus-responsive drug delivery fields.

Investigation of mechanical properties and structural integrity of graphene aerogels via molecular dynamics simulations.

Graphene aerogel (GA), a 3D carbon-based nanostructure built on 2D graphene sheets, is well known for being the lightest solid material ever synthesized. It also possesses many other exceptional properties, such as high specific surface area and large liquid absorption capacity, thanks to its ultra-high porosity. Computationally, the mechanical properties of GA have been studied by molecular dynamics (MD) simulations, which uncover nanoscale mechanisms beyond experimental observations. However, studies on how GA structures and properties evolve in response to simulation parameter changes, which provide valuable insights to experimentalists, have been lacking. In addition, the differences between the calculated properties via simulations and experimental measurements have rarely been discussed. To address the shortcomings mentioned above, in this study, we systematically study various mechanical properties and the structural integrity of GA as a function of a wide range of simulation parameters. Results show that during the in silico GA preparation, smaller and less spherical inclusions (mimicking the effect of water clusters in experiments) are conducive to strength and stiffness but may lead to brittleness. Additionally, it is revealed that a structurally valid GA in the MD simulation requires the number of bonds per atom to be at least 1.40, otherwise the GA building blocks are not fully interconnected. Finally, our calculation results are compared with experiments to showcase both the power and the limitations of the simulation technique. This work may shed light on the improvement of computational approaches for GA as well as other novel nanomaterials.

Combination therapy with itraconazole and terbinafine for phaeohyphomycosis caused by Exophiala spinifera: A case report and literature review.

Exophiala spinifera is a rare dematiaceous fungus causing cutaneous, subcutaneous and disseminated phaeohyphomycosis (PHM). Standard antifungal therapy for PHM is still uncertain. Here, we report a case of a Chinese male with PHM caused by E. spinifera, who received significant clinical improvement after the treatment with oral itraconazole and terbinafine. With the aim of evaluating the antifungal therapy for PHM caused by E. spinifera, a detailed review was performed.

Clinicopathological characteristics and prognostic factors in axial chondroblastomas: a retrospective analysis of 61 cases and comparison with extra-axial chondroblastomas.

BACKGROUND: A comprehensive understanding of the clinical characteristics and prognostic factors associated with axial chondroblastoma (ACB) is still lacking. This study aimed to understand the clinical characteristics and prognostic factors of axial chondroblastoma (ACB) and compare them with extra-axial chondroblastoma (EACB). METHODS: A retrospective review of our institution's local database was conducted, encompassing a total of 132 CB patients, of which 61 were diagnosed with ACB and 71 with EACB. Immunohistochemistry was employed to evaluate the expression levels of vimentin, S100, and cytokeratin. RESULTS: ACB and EACB shared similar characteristics, with the exception of advanced age, tumor size, elevated Vim expression, incidence of surrounding tissue invasion, and postoperative sensory or motor dysfunction. While wide resection and absence of surrounding tissue invasion consistently showed a favorable association with survival in both ACB and EACB cohorts during univariate analysis, most parameters exhibited differential prognostic significance between the two groups. Notably, the significant prognostic factors for local recurrence-free survival in the ACB cohort included the type of resection and the presence of chicken-wire calcification. In the multivariate analysis of overall survival, the type of resection emerged as a significant predictor in the ACB cohort, whereas in the EACB group, the type of resection and the occurrence of postoperative sensory or motor dysfunction were predictive of overall survival. CONCLUSION: There may exist distinct biological behaviors between ACB and EACB, thereby providing valuable insights into the prognostic characteristics of ACB patients and contributing to enhanced outcome prediction in this particular patient population.

Improved UAV-to-Ground Multi-Target Tracking Algorithm Based on StrongSORT.

Unmanned aerial vehicles (UAV) are essential for aerial reconnaissance and monitoring. One of the greatest challenges facing UAVs is vision-based multi-target tracking. Multi-target tracking algorithms that depend on visual data are utilized in a variety of fields. In this study, we present a comprehensive framework for real-time tracking of ground robots in forest and grassland environments. This framework utilizes the YOLOv5n detection algorithm and a multi-target tracking algorithm for monitoring ground robot activities in real-time video streams. We optimized both detection and re-identification networks to enhance real-time target detection. The StrongSORT tracking algorithm was selected carefully to alleviate the loss of tracked objects due to factors like camera jitter, intersecting and overlapping targets, and smaller target sizes. The YOLOv5n algorithm was used to train the dataset, and the StrongSORT tracking algorithm incorporated the best-trained model weights. The algorithm's performance has greatly improved, as demonstrated by experimental results. The number of ID switches (IDSW) has decreased by sixfold, IDF1 has increased by 7.93%, and false positives (FP) have decreased by 30.28%. Additionally, the tracking speed has reached 38 frames per second. These findings validate our algorithm's ability to fulfill real-time tracking requisites on UAV platforms, delivering dependable resolutions for dynamic multi-target tracking on land.

Deep Learning-Based Multiclass Brain Tissue Segmentation in Fetal MRIs.

Fetal brain tissue segmentation is essential for quantifying the presence of congenital disorders in the developing fetus. Manual segmentation of fetal brain tissue is cumbersome and time-consuming, so using an automatic segmentation method can greatly simplify the process. In addition, the fetal brain undergoes a variety of changes throughout pregnancy, such as increased brain volume, neuronal migration, and synaptogenesis. In this case, the contrast between tissues, especially between gray matter and white matter, constantly changes throughout pregnancy, increasing the complexity and difficulty of our segmentation. To reduce the burden of manual refinement of segmentation, we proposed a new deep learning-based segmentation method. Our approach utilized a novel attentional structural block, the contextual transformer block (CoT-Block), which was applied in the backbone network model of the encoder-decoder to guide the learning of dynamic attentional matrices and enhance image feature extraction. Additionally, in the last layer of the decoder, we introduced a hybrid dilated convolution module, which can expand the receptive field and retain detailed spatial information, effectively extracting the global contextual information in fetal brain MRI. We quantitatively evaluated our method according to several performance measures: dice, precision, sensitivity, and specificity. In 80 fetal brain MRI scans with gestational ages ranging from 20 to 35 weeks, we obtained an average Dice similarity coefficient (DSC) of 83.79%, an average Volume Similarity (VS) of 84.84%, and an average Hausdorff95 Distance (HD95) of 35.66 mm. We also used several advanced deep learning segmentation models for comparison under equivalent conditions, and the results showed that our method was superior to other methods and exhibited an excellent segmentation performance.

Underwater Fish Segmentation Algorithm Based on Improved PSPNet Network.

With the sustainable development of intelligent fisheries, accurate underwater fish segmentation is a key step toward intelligently obtaining fish morphology data. However, the blurred, distorted and low-contrast features of fish images in underwater scenes affect the improvement in fish segmentation accuracy. To solve these problems, this paper proposes a method of underwater fish segmentation based on an improved PSPNet network (IST-PSPNet). First, in the feature extraction stage, to fully perceive features and context information of different scales, we propose an iterative attention feature fusion mechanism, which realizes the depth mining of fish features of different scales and the full perception of context information. Then, a SoftPool pooling method based on fast index weighted activation is used to reduce the numbers of parameters and computations while retaining more feature information, which improves segmentation accuracy and efficiency. Finally, a triad attention mechanism module, triplet attention (TA), is added to the different scale features in the golden tower pool module so that the space attention can focus more on the specific position of the fish body features in the channel through cross-dimensional interaction to suppress the fuzzy distortion caused by background interference in underwater scenes. Additionally, the parameter-sharing strategy is used in this process to make different scale features share the same learning weight parameters and further reduce the numbers of parameters and calculations. The experimental results show that the method presented in this paper yielded better results for the DeepFish underwater fish image dataset than other methods, with 91.56% for the Miou, 46.68 M for Params and 40.27 G for GFLOPS. In the underwater fish segmentation task, the method improved the segmentation accuracy of fish with similar colors and water quality backgrounds, improved fuzziness and small size and made the edge location of fish clearer.