Reviving Static Charts into Live Charts.

Data charts are prevalent across various fields due to their efficacy in conveying complex data relationships. However, static charts may sometimes struggle to engage readers and efficiently present intricate information, potentially resulting in limited understanding. We introduce "Live Charts," a new format of presentation that decomposes complex information within a chart and explains the information pieces sequentially through rich animations and accompanying audio narration. We propose an automated approach to revive static charts into Live Charts. Our method integrates GNN-based techniques to analyze the chart components and extract data from charts. Then we adopt large natural language models to generate appropriate animated visuals along with a voice-over to produce Live Charts from static ones. We conducted a thorough evaluation of our approach, which involved the model performance, use cases, a crowd-sourced user study, and expert interviews. The results demonstrate Live Charts offer a multi-sensory experience where readers can follow the information and understand the data insights better. We analyze the benefits and drawbacks of Live Charts over static charts as a new information consumption experience.

Hierarchical Recognizing Vector Graphics and A New Chart-based Vector Graphics Dataset.

The conventional approach to image recognition has been based on raster graphics, which can suffer from aliasing and information loss when scaled up or down. In this paper, we propose a novel approach that leverages the benefits of vector graphics for object localization and classification. Our method, called YOLaT (You Only Look at Text), takes the textual document of vector graphics as input, rather than rendering it into pixels. YOLaT builds multi-graphs to model the structural and spatial information in vector graphics and utilizes a dual-stream graph neural network (GNN) to detect objects from the graph. However, for real-world vector graphics, YOLaT only models in flat GNN with vertexes as nodes ignore higher-level information of vector data. Therefore, we propose YOLaT++ to learn Multi-level Abstraction Feature Learning from a new perspective: Primitive Shapes to Curves and Points. On the other hand, given few public datasets focus on vector graphics, data-driven learning cannot exert its full power on this format. We provide a large-scale and challenging dataset for Chart-based Vector Graphics Detection and Chart Understanding, termed VG-DCU, with vector graphics, raster graphics, annotations, and raw data drawn for creating these vector charts. Experiments show that the YOLaT series outperforms both vector graphics and raster graphics-based object detection methods on both subsets of VG-DCU in terms of both accuracy and efficiency, showcasing the potential of vector graphics for image recognition tasks. Our codes, models, and the VG-DCU dataset are available at: https://github.com/microsoft/YOLaT-VectorGraphicsRecognition.

Online Streaming Video Super-Resolution With Convolutional Look-Up Table.

Online video streaming has fundamental limitations on the transmission bandwidth and computational capacity and super-resolution is a promising potential solution. However, applying existing video super-resolution methods to online streaming is non-trivial. Existing video codecs and streaming protocols (e.g., WebRTC) dynamically change the video quality both spatially and temporally, which leads to diverse and dynamic degradations. Furthermore, online streaming has a strict requirement for latency that most existing methods are less applicable. As a result, this paper focuses on the rarely exploited problem setting of online streaming video super resolution. To facilitate the research on this problem, a new benchmark dataset named LDV-WebRTC is constructed based on a real-world online streaming system. Leveraging the new benchmark dataset, we propose a novel method specifically for online video streaming, which contains a convolution and Look-Up Table (LUT) hybrid model to achieve better performance-latency trade-off. To tackle the changing degradations, we propose a mixture-of-expert-LUT module, where a set of LUT specialized in different degradations are built and adaptively combined to handle different degradations. Experiments show our method achieves 720P video SR around 100 FPS, while significantly outperforms existing LUT-based methods and offers competitive performance compared to efficient CNN-based methods. Code is available at https://github.com/quzefan/ConvLUT.

Lacrimal Gland Adenocarcinoma Clinicopathologic Features and Outcomes Compared With Those of Lacrimal Gland Adenoid Cystic Carcinoma.

PURPOSE: Lacrimal gland (LG) adenocarcinomas (ACs) are rare, with limited data. We compared clinicopathologic features and local recurrence, distant metastasis, and survival rates between LG AC and LG adenoid cystic carcinoma (ACC). METHODS: The records of LG AC patients treated from 2008 to 2022 and LG ACC patients treated from 1998 to 2022 at the same center were retrospectively reviewed. RESULTS: The study included 20 patients with AC; 10 de-novo AC, 10 ex-pleomorphic AC; and 51 ACC patients. The median age at diagnosis was 61 years for de-novo AC, 54 years for ex-pleomorphic AC, and 45 years for ACC. All groups had male predominance. The initial T category was T2 in 50% (5/10) of de-novo ACs; 60% (6/10) of ex-pleomorphic ACs; and 59% (30/51) of ACCs. Perineural invasion was present in 33% (5/15) of ACs and 90% (45/50) of ACCs ( p < 0.001). Of the 20 AC patients, 14 had eye-sparing surgery; 4 had orbital exenteration; and 2 had unresectable disease. All AC patients received postoperative radiotherapy and 15 (75%) received concurrent chemotherapy. Fourteen AC patients were tested for human growth factor receptor 2 expression, and 10 (71%) were human growth factor receptor 2 positive; 5 received human growth factor receptor 2-targeted therapy. AC and ACC had similar 5-year recurrence rates (20% and 33%, respectively, p = 0.31) and metastasis rates (20% and 34%, respectively, p = 0.30). de-novo AC, ex-pleomorphic AC, and ACC had similar 5-year disease-specific survival rates (80%, 79%, and 81%, respectively, p > 0.99). CONCLUSIONS: LG AC and ACC have similar baseline clinicopathologic features, except that perineural invasion is more common in ACC, and similar recurrence, metastasis, and survival rates. Human growth factor receptor 2-targeted therapy may be appropriate in some patients with LG AC.

Learning Domain Invariant Prompt for Vision-Language Models.

Prompt learning stands out as one of the most efficient approaches for adapting powerful vision-language foundational models like CLIP to downstream datasets by tuning learnable prompt vectors with very few samples. However, despite its success in achieving remarkable performance on in-domain data, prompt learning still faces the significant challenge of effectively generalizing to novel classes and domains. Some existing methods address this concern by dynamically generating distinct prompts for different domains. Yet, they overlook the inherent potential of prompts to generalize across unseen domains. To address these limitations, our study introduces an innovative prompt learning paradigm, called MetaPrompt, aiming to directly learn domain invariant prompt in few-shot scenarios. To facilitate learning prompts for image and text inputs independently, we present a dual-modality prompt tuning network comprising two pairs of coupled encoders. Our study centers on an alternate episodic training algorithm to enrich the generalization capacity of the learned prompts. In contrast to traditional episodic training algorithms, our approach incorporates both in-domain updates and domain-split updates in a batch-wise manner. For in-domain updates, we introduce a novel asymmetric contrastive learning paradigm, where representations from the pre-trained encoder assume supervision to regularize prompts from the prompted encoder. To enhance performance on out-of-domain distribution, we propose a domain-split optimization on visual prompts for cross-domain tasks or textual prompts for cross-class tasks during domain-split updates. Extensive experiments across 11 datasets for base-to-new generalization and 4 datasets for domain generalization exhibit favorable performance. Compared with the state-of-the-art method, MetaPrompt achieves an absolute gain of 1.02% on the overall harmonic mean in base-to-new generalization and consistently demonstrates superiority over all benchmarks in domain generalization.

Addressing subject heterogeneity in time-dependent discrimination for biomarker evaluation.

Accurate discrimination has been the central goal in identifying biomarkers for monitoring disease progression and early detection. Acknowledging the fact that discrimination accuracy of biomarkers for a time-to-event outcome often changes over time, local measures such as the time-dependent receiver operating characteristic curve and its area under the curve (AUC) are used to assess time-dependent predictive discrimination. However, such measures do not address subject heterogeneity, although the impact of covariates including demographics, disease-related characteristics, and other clinical information on the discriminatory performance of biomarkers needs to be investigated before their clinical use. We propose the covariate-specific time-dependent AUC, a measure for covariate-adjusted discrimination. We develop a regression model on the covariate-specific time-dependent AUC to understand how and in what magnitude the covariates influence biomarker performance. Then we construct a pseudo partial-likelihood for estimation and inference. This is followed by our establishing the asymptotic properties of the proposed estimators and provide variance estimation. The simulation studies and application to the AIDS Clinical Trials Group 175 data demonstrate that the proposed method offers an informative tool for inferring covariate-specific and time-dependent predictive discrimination.

Orbital and periocular complications in patients with sinonasal tumours with orbital invasion.

AIMS: The purpose of this study was to determine the frequency and associated risk factors of orbital/periocular complications in patients with sinonasal tumour with orbital invasion managed with eye-sparing treatments. METHODS: A retrospective case series of patients with primary sinonasal tumour with orbital invasion from January 2008 to December 2018. Patient factors were compared between the following groups: (1)patients with orbital/periocular complications versus those who did not and (2) patients who needed secondary oculoplastic surgical procedures versus those who did not. RESULTS: Out of 80 patients, 48 had eye-sparing surgery, 8 had orbital exenteration and 24 were managed non-surgically. The most common histology was squamous cell carcinoma (n=28, 35%). Among the eye-sparing treatment group, 51/72 patients experienced one or more orbital/periocular complication(s), with motility deficit (N=26, 36%) being the most frequent. Factors associated with higher risk of complications included tumour involving the orbital floor (p=0.019), clinical disease stage III/IV (p=0.038), maxillectomy (p=0.004), resection of the orbital floor (p=0.027) and cigarette smoking (p=0.041). Tumour involving the orbital floor had an OR of 3.9 (95% CI 1.3 to 11.6, p=0.016) in predicting orbital/periocular complication. In the eye-sparing surgery group, the most frequent secondary oculoplastic procedures was dacryocystorhinostomy (n=6, 13%). The use of a free flap in reconstruction had an OR of 8.2 (95% CI 2.1 to 31.8, p=0.002) in predicting need for secondary oculoplastic surgery. CONCLUSION: Majority of patients with sinonasal tumours and secondary orbital invasion were managed with eye-sparing multidisciplinary treatments. Preservation of the eye can lead to reasonably good functional outcome despite expected orbital and periocular complications.

Evaluating dynamic and predictive discrimination for recurrent event models: use of a time-dependent C-index.

Interest in analyzing recurrent event data has increased over the past few decades. One essential aspect of a risk prediction model for recurrent event data is to accurately distinguish individuals with different risks of developing a recurrent event. Although the concordance index (C-index) effectively evaluates the overall discriminative ability of a regression model for recurrent event data, a local measure is also desirable to capture dynamic performance of the regression model over time. Therefore, in this study, we propose a time-dependent C-index measure for inferring the model's discriminative ability locally. We formulated the C-index as a function of time using a flexible parametric model and constructed a concordance-based likelihood for estimation and inference. We adapted a perturbation-resampling procedure for variance estimation. Extensive simulations were conducted to investigate the proposed time-dependent C-index's finite-sample performance and estimation procedure. We applied the time-dependent C-index to three regression models of a study of re-hospitalization in patients with colorectal cancer to evaluate the models' discriminative capability.

Content-Adaptive Auto-Occlusion Network for Occluded Person Re-Identification.

The occluded person re-identification (ReID) aims to match person images captured in severely occluded environments. Current occluded ReID works mostly rely on auxiliary models or employ a part-to-part matching strategy. However, these methods may be sub-optimal since the auxiliary models are constrained by occlusion scenes and the matching strategy will deteriorate when both query and gallery set contain occlusion. Some methods attempt to solve this problem by applying image occlusion augmentation (OA) and have shown great superiority in their effectiveness and lightness. But there are two defects that existed in the previous OA-based method: 1) The occlusion policy is fixed throughout the entire training and cannot be dynamically adjusted based on the current training status of the ReID network. 2) The position and area of the applied OA are completely random, without reference to the image content to choose the most suitable policy. To address these challenges, we propose a novel Content-Adaptive Auto-Occlusion Network (CAAO), that is able to dynamically select the proper occlusion region of an image based on its content and the current training status. Specifically, CAAO consists of two parts: the ReID network and the Auto-Occlusion Controller (AOC) module. AOC automatically generates the optimal OA policy based on the feature map extracted from the ReID network and applies occlusion on the images for ReID network training. An on-policy reinforcement learning based alternating training paradigm is proposed to iteratively update the ReID network and AOC module. Comprehensive experiments on occluded and holistic person ReID benchmarks demonstrate the superiority of CAAO.

Trichorhinophalangeal syndrome type 1 (TRPS1) expression in male breast carcinoma.

Currently, there is a paucity of highly specific and sensitive markers to identify breast carcinoma in male patients. Immunohistochemical stains commonly used for unmasking primary breast carcinomas include estrogen receptor (ER) and GATA3. However, these markers are commonly expressed in carcinomas originating from other organ systems and can be reduced in breast carcinomas with higher histologic grades. Androgen receptor (AR) may be used to highlight primary male breast cancer, but this marker can also be expressed in other carcinomas. We evaluated TRPS1, a highly sensitive and specific marker for female breast carcinoma, in cases of male breast carcinoma. Through an institutional database search, we identified 72 cases of primary invasive breast carcinoma in male patients. Among ER/progesterone receptor (PR)-positive cancers, 97% showed intermediate or high positivity for both TRPS1 and GATA3. Among HER2-positive cancers, 100% showed intermediate or high positivity for TRPS1 and GATA3. One case of triple-negative breast cancer was collected, showing high positivity for TRPS1 and negativity for GATA3. AR staining was non-specific and heterogeneous: 76% showed high positivity, but the remaining 24% showed low or intermediate positivity. Additionally, among 29 cases of metastatic carcinoma to male breast tissue, 93% were negative for TRPS1, and the remaining 2 cases (7%), which were carcinomas from salivary gland primary tumors, were intermediate positive. TRPS1 is a sensitive and specific marker in the unmasking of male primary invasive breast carcinoma across different subtypes. Additionally, TRPS1 is not expressed in metastatic carcinomas of multiple primaries, with the exception of salivary gland primaries.

Safety and immunogenicity of heterologous boosting with orally aerosolised or intramuscular Ad5-nCoV vaccine and homologous boosting with inactivated vaccines (BBIBP-CorV or CoronaVac) in children and adolescents: a randomised, open-label, parallel-controlled, non-inferiority, single-centre study.

BACKGROUND: Heterologous booster immunisation with orally administered aerosolised Ad5-nCoV vaccine (AAd5) has been shown to be safe and highly immunogenic in adults. Here, we aimed to assess the safety and immunogenicity of heterologous booster immunisation with orally administered AAd5 in children and adolescents aged 6-17 years who had received two doses of inactivated vaccine (BBIBP-CorV or CoronaVac). METHODS: We did a randomised, open-label, parallel-controlled, non-inferiority study to assess the safety and immunogenicity of heterologous booster immunisation with AAd5 (0·1 mL) or intramuscular Ad5-nCoV vaccine (IMAd5; 0·3 mL) and homologous booster immunisation with inactivated vaccine (BBIBP-CorV or CoronaVac; 0·5 mL) in children (aged 6-12 years) and adolescents (aged 13-17 years) who had received two doses of inactivated vaccine at least 3 months earlier in Hunan, China. Children and adolescents who were previously immunised with two-dose BBIBP-CorV or CoronaVac were recruited for eligibility screening at least 3 months after the second dose. A stratified block method was used for randomisation, and participants were stratified by age and randomly assigned (3:1:1) to receive AAd5, IMAd5, or inactivated vaccine. The study staff and participants were not masked to treatment allocation. Laboratory and statistical staff were masked during the study. In this interim analysis, adverse events within 14 days and geometric mean titre (GMT) of serum neutralising antibodies on day 28 after the booster vaccination, based on the per-protocol population, were used as the primary outcomes. The analysis of non-inferiority was based on comparison using a one-sided 97·5% CI with a non-inferiority margin of 0·67. This study was registered at ClinicalTrials.gov, NCT05330871, and is ongoing. FINDINGS: Between April 17 and May 28, 2022, 436 participants were screened and 360 were enrolled: 220 received AAd5, 70 received IMAd5, and 70 received inactivated vaccine. Within 14 days after booster vaccination, vaccine-related adverse reactions were reported: 35 adverse events (in 13 [12%] of 110 children and 22 [20%] of 110 adolescents) in 220 individuals in the AAd5 group, 35 (in 18 [51%] of 35 children and 17 [49%] of 35 adolescents) in 70 individuals in the IMAd5 group, and 13 (in five [14%] of 35 children and eight [23%] of 35 adolescents) in 70 individuals in the inactivated vaccine group. Solicited adverse reactions were also reported: 34 (13 [12%] of 110 children and 21 [10%] of 110 adolescents) in 220 individuals in the AAd5 group, 34 (17 [49%] of 35 children and 17 [49%] of 35 adolescents) in 70 individuals in the IMAd5 group, and 12 (five [14%] of 35 children and seven [20%] of 35 adolescents) in 70 individuals in the inactivated vaccine group. The GMTs of neutralising antibodies against ancestral SARS-CoV-2 Wuhan-Hu-1 (Pango lineage B) in the AAd5 group were significantly higher than the GMTs in the inactivated vaccine group (adjusted GMT ratio 10·2 [95% CI 8·0-13·1]; p<0·0001). INTERPRETATION: Our study shows that a heterologous booster with AAd5 is safe and highly immunogenic against ancestral SARS-CoV-2 Wuhan-Hu-1 in children and adolescents. FUNDING: National Key R&D Program of China.

The reason and mechanism of propylene glycol alginate sodium sulfate (PSS) mediated allergic side effect.

Propylene glycol alginate sodium sulfate (PSS) is a heparinoid polysaccharide drug used in clinic for >30 years in China. But its allergy events happened from time to time and should not be ignored. Here, ammonium salt in PSS (PSS-NH4+), PSS fractions with high Mw (PSS-H-Mw) and low mannuronic acid (M) to guluronic acid (G) ratio (PSS-L-M/G) were found to induce allergic response by the structure-activity and impurity-activity relationships in vitro. Furthermore, we confirmed the reason and elucidated the mechanism accounted for allergic side effect of PSS in vivo. It was found that high IgE levels in PSS-NH4+ and PSS-H-Mw groups upregulate the cascade expression of Lyn-Syk-Akt or Erk and second messenger Ca2+, which accelerated mast cells (MCs) degranulation to release histamine, LTB4, TPS, and finally induced lung tissue injury. PSS-L-M/G caused a mild allergic symptom because it only enhanced the expression of p-Lyn and histamine release. In brief, PSS-NH4+ and PSS-H-Mw were main reasons to result in allergic response. Our results suggested that it is very necessary to control the range of Mw and the content of impurities (< 1 % ammonium salt) of PSS to guarantee its safety and effectiveness in clinical treatment.

Ethnic-specificity, evolution origin and deleteriousness of Asian BRCA variation revealed by over 7500 BRCA variants derived from Asian population.

Pathogenic variation in BRCA1 and BRCA2 (BRCA) causes high risk of breast and ovarian cancer, and BRCA variation data are important markers for BRCA-related clinical cancer applications. However, comprehensive BRCA variation data are lacking from the Asian population despite its large population size, heterogenous genetic background and diversified living environment across the Asia continent. We performed a systematic study on BRCA variation in Asian population including extensive data mining, standardization, annotation and characterization. We identified 7587 BRCA variants from 685 592 Asian individuals in 40 Asia countries and regions, including 1762 clinically actionable pathogenic variants and 4915 functionally unknown variants (https://genemutation.fhs.um.edu.mo/Asian-BRCA/). We observed the highly ethnic-specific nature of Asian BRCA variants between Asian and non-Asian populations and within Asian populations, highlighting that the current European descendant population-based BRCA data is inadequate to reflect BRCA variation in the Asian population. We also provided archeological evidence for the evolutionary origin and arising time of Asian BRCA variation. We further provided structural-based evidence for the deleterious variants enriched within the functionally unknown Asian BRCA variants. The data from our study provide a current view of BRCA variation in the Asian population and a rich resource to guide clinical applications of BRCA-related cancer for the Asian population.

Visual outcome and ocular complications in patients with lacrimal gland carcinoma after eye-sparing surgery and adjuvant radiation therapy.

BACKGROUND: We report visual outcomes and ocular complications in patients with lacrimal gland carcinoma who had eye-sparing surgery followed by radiotherapy. METHODS: This review included consecutive patients with lacrimal gland carcinoma who underwent eye-sparing surgery and adjuvant radiotherapy or concurrent chemoradiation therapy between 2007 and 2018. Clinical data, including details of ophthalmological examinations and radiation treatment were reviewed. RESULTS: The study included 23 patients, 15 males and 8 females, with median age 51 years. Twenty patients (87%) received intensity-modulated proton therapy; 3 (13%) received intensity-modulated radiotherapy. Nineteen patients (83%) received concurrent chemotherapy. After a median follow-up time of 37 months (range: 8-83), 13 patients (57%) had best-corrected visual acuity 20/40 or better, 3 (13%) had moderate vision loss (between 20/40 and 20/200) and 7 (30%) had severe vision loss (20/200 or worse). The most common ocular complications were dry eye disease (21 patients; 91%), radiation retinopathy (16; 70%) and cataract progression (11; 49%). Tumour crossing the orbital midline (p=0.014) and Hispanic ethnicity (p=0.014) were associated with increased risk of severe vision loss. The risk of radiation retinopathy was significantly different among the three racial groups; Hispanic patients (n=3) had the highest rate of retinopathy (p<0.001). Tumour size, initial T category and total prescribed radiation dose were not significantly associated with severe vision loss. CONCLUSION: Eye-sparing surgery followed by adjuvant radiotherapy in patients with lacrimal gland carcinoma has a reasonable overall visual prognosis. Patients with tumours crossing the orbital midline and Hispanic patients have a higher risk of severe vision loss.

Synchronous Extraction, Antioxidant Activity Evaluation, and Composition Analysis of Carbohydrates and Polyphenols Present in Artichoke Bud.

This study investigated the optimization of ultrasonic-assisted aqueous two-phase synchronous extraction of carbohydrates and polyphenols present in artichoke bud, evaluated their antioxidant activities in vitro, and analyzed the composition of carbohydrates and polyphenols by high-performance liquid chromatography (HPLC). The powder mass, ultrasonic time, ammonium sulfate concentration, and alcohol-water ratio were considered the influencing factors based on the single-factor experiment results, and a dual-response surface model was designed to optimize the synchronous extraction process to extract carbohydrates and polyphenols. The antioxidant activity was evaluated by measuring the scavenging capacity of ABTS+· and DPPH· and the reducing capacity of Fe3+. The optimal process conditions in this study were as follows: the powder mass of 1.4 g, ammonium sulfate concentration of 0.34 g/mL, alcohol-water ratio of 0.4, and ultrasonic time of 43 min. The polyphenol content in artichoke bud was 5.32 ± 0.13 mg/g, and the polysaccharide content was 74.78 ± 0.11 mg/g. An experiment on in vitro antioxidant activity showed that both carbohydrates and polyphenols had strong antioxidant activities, and the antioxidant activity of polyphenols was stronger than that of carbohydrates. The HPLC analysis revealed that the carbohydrates in artichoke bud were mannose, rhamnose, glucuronic acid, galacturonic acid, glucose, galactose, and arabinose, and the molar ratio was 10.77:25.22:2.37:15.74:125.39:48.62:34.70. The polyphenols comprised chlorogenic acid, 4-dicaffeoylquinic acid, caffeic acid, 1,3-dicaffeoylqunic acid, isochlorogenic acid B, isochlorogenic acid A, cynarin, and isochlorogenic acid C, and the contents were 0.503, 0.029, 0.022, 0.017, 0.008, 0.162, 1.621, 0.030 mg/g, respectively. This study also showed that the carbohydrates and polyphenols in artichoke bud could be important natural antioxidants, and the composition analysis of HPLC provided directions for their future research. Carbohydrates and polyphenols in artichoke buds can be separated and enriched using the optimized process technology, and it is an effective means of extracting ingredients from plants.

Conditional Feature Learning Based Transformer for Text-Based Person Search.

Text-based person search aims at retrieving the target person in an image gallery using a descriptive sentence of that person. The core of this task is to calculate a similarity score between the pedestrian image and description, which requires inferring the complex latent correspondence between image sub-regions and textual phrases at different scales. Transformer is an intuitive way to model the complex alignment by its self-attention mechanism. Most previous Transformer-based methods simply concatenate image region features and text features as input and learn a cross-modal representation in a brute force manner. Such weakly supervised learning approaches fail to explicitly build alignment between image region features and text features, causing an inferior feature distribution. In this paper, we present CFLT, Conditional Feature Learning based Transformer. It maps the sub-regions and phrases into a unified latent space and explicitly aligns them by constructing conditional embeddings where the feature of data from one modality is dynamically adjusted based on the data from the other modality. The output of our CFLT is a set of similarity scores for each sub-region or phrase rather than a cross-modal representation. Furthermore, we propose a simple and effective multi-modal re-ranking method named Re-ranking scheme by Visual Conditional Feature (RVCF). Benefit from the visual conditional feature and better feature distribution in our CFLT, the proposed RVCF achieves significant performance improvement. Experimental results show that our CFLT outperforms the state-of-the-art methods by 7.03% in terms of top-1 accuracy and 5.01% in terms of top-5 accuracy on the text-based person search dataset.

Conditional Feature Embedding by Visual Clue Correspondence Graph for Person Re-Identification.

Although Person Re-Identification has made impressive progress, difficult cases like occlusion, change of view-point, and similar clothing still bring great challenges. In order to tackle these challenges, extracting discriminative feature representation is crucial. Most of the existing methods focus on extracting ReID features from individual images separately. However, when matching two images, we propose that the ReID features of a query image should be dynamically adjusted based on the contextual information from the gallery image it matches. We call this type of ReID features conditional feature embedding. In this paper, we propose a novel ReID framework that extracts conditional feature embedding based on the aligned visual clues between image pairs, called Clue Alignment based Conditional Embedding (CACE-Net). CACE-Net applies an attention module to build a detailed correspondence graph between crucial visual clues in image pairs and uses discrepancy-based GCN to embed the obtained complex correspondence information into the conditional features. The experiments show that CACE-Net achieves state-of-the-art performance on three public datasets.

Garlic polysaccharides: A review on their extraction, isolation, structural characteristics, and bioactivities.

Garlic has been considered as a source of highly promising functional food and traditional herbal medicine for thousands of years. Garlic polysaccharides is one of the important effective components of garlic, which has various bioactivities, including immune-enhancing, hepatoprotective, and antioxidant. Garlic polysaccharides is mainly composed of monosaccharides, such as Fru, Glc, and Gal, having a (2 â†’ 1)-linked ß-D-Fruf backbone with (2 â†’ 6)-linked ß-D-Fruf side chains. With great marketing potential and development prospects, garlic polysaccharides has drawn much attention from researchers worldwide. Therefore, this review aimed at providing systematic and current information on the extraction, isolation, structural characteristics, and bioactivities of garlic polysaccharides to support their further application as therapeutic agents and functional foods.

Human Co-Parsing Guided Alignment for Occluded Person Re-identification.

Occluded person re-identification (ReID) is a challenging task due to more background noises and incomplete foreground information. Although existing human parsing-based ReID methods can tackle this problem with semantic alignment at the finest pixel level, their performance is heavily affected by the human parsing model. Most supervised methods propose to train an extra human parsing model aside from the ReID model with cross-domain human parts annotation, suffering from expensive annotation cost and domain gap; Unsupervised methods integrate a feature clustering-based human parsing process into the ReID model, but lacking supervision signals brings less satisfactory segmentation results. In this paper, we argue that the pre-existing information in the ReID training dataset can be directly used as supervision signals to train the human parsing model without any extra annotation. By integrating a weakly supervised human co-parsing network into the ReID network, we propose a novel framework that exploits shared information across different images of the same pedestrian, called the Human Co-parsing Guided Alignment (HCGA) framework. Specifically, the human co-parsing network is weakly supervised by three consistency criteria, namely global semantics, local space, and background. By feeding the semantic information and deep features from the person ReID network into the guided alignment module, features of the foreground and human parts can then be obtained for effective occluded person ReID. Experiment results on two occluded and two holistic datasets demonstrate the superiority of our method. Especially on Occluded-DukeMTMC, it achieves 70.2% Rank-1 accuracy and 57.5% mAP.

Spatial Accessibility Assessment of Prehospital EMS with a Focus on the Elderly Population: A Case Study in Ningbo, China.

The spatial accessibility of prehospital EMS is particularly important for the elderly population's physiological functions. Due to the recent expansion of aging populations all over the globe, elderly people's spatial accessibility to prehospital EMS presents a serious challenge. An efficient strategy to address this issue involves using geographic information systems (GIS)-based tools to evaluate the spatial accessibility in conjunction with the spatial distribution of aging people, available road networks, and prehospital EMS facilities. This study employed gravity model and empirical Bayesian Kriging (EBK) interpolation analysis to evaluate the elderly's spatial access to prehospital EMS in Ningbo, China. In our study, we aimed to solve the following specific research questions: In the study area, "what are the characteristics of the prehospital EMS demand of the elderly?" "Do the elderly have equal and convenient spatial access to prehospital EMS?" and "How can we satisfy the prehospital EMS demand of an aging population, improve their spatial access to prehospital EMS, and then ensure their quality of life?" The results showed that 37.44% of patients admitted to prehospital EMS in 2020 were 65 years and older. The rate of utilization of ambulance services by the elderly was 27.39 per 1000 elderly residents. Ambulance use by the elderly was the highest in the winter months and the lowest in the spring months (25.90% vs. 22.38%). As for the disease spectrum, the main disease was found to be trauma and intoxication (23.70%). The mean accessibility score was only 1.43 and nearly 70% of demand points had scored lower than 1. The elderly's spatial accessibility to prehospital EMS had a central-outward gradient decreasing trend from the central region to the southeast and southwest of the study area. Our proposed methodology and its spatial equilibrium results could be taken as a benchmark of prehospital care capacity and help inform authorities' efforts to develop efficient, aging-focused spatial accessibility plans.