Explore the vast range of titles available.

Object detection, as a fundamental task in computer vision, has been developed enormously, but is still challenging work, especially for Unmanned Aerial Vehicle (UAV) perspective due to small scale of the target. In this study, the authors develop a special detection method for small objects in UAV perspective. Based on YOLOv3, the Resblock in darknet is first optimized by concatenating two ResNet units that have the same width and height. Then, the entire darknet structure is improved by increasing convolution operation at an early layer to enrich spatial information. Both these two optimizations can enlarge the receptive filed. Furthermore, UAV-viewed dataset is collected to UAV perspective or small object detection. An optimized training method is also proposed based on collected UAV-viewed dataset. The experimental results on public dataset and our collected UAV-viewed dataset show distinct performance improvement on small object detection with keeping the same level performance on normal dataset, which means our proposed method adapts to different kinds of conditions.

Background The application of multiparametric magnetic resonance imaging (mpMRI) for diagnosis of prostate cancer has been recommended by the European Association of Urology (EAU), National Comprehensive Cancer Network (NCCN), and European Society of Urogenital Radiology (ESUR) guidelines. The purpose of this study is to systematically review the literature on assessing the accuracy of mpMRI in patients with suspicion of prostate cancer. Method We searched Embase, Pubmed and Cochrane online databases from January 12,000 to October 272,018 to extract articles exploring the possibilities that the pre-biopsy mpMRI can enhance the diagnosis accuracy of prostate cancer. The numbers of true- and false-negative results and true- and false-positive ones were extracted to calculate the corresponding sensitivity and specificity of mpMRI. Study quality was assessed using QUADAS-2 tool. Random effects meta-analysis and a hierarchical summary receiver operating characteristic (HSROC) plot were performed for further study. Results After searching, we acquired 3741 articles for reference, of which 29 studies with 8503 participants were eligible for inclusion. MpMRI maintained impressive diagnostic value, the area under the HSROC curve was 0.87 (95%CI,0.84–0.90). The sensitivity and specificity for mpMRI were 0.87 [95%CI, 0.81–0.91] and 0.68 [95%CI,0.56–0.79] respectively. The positive likelihood ratio was 2.73 [95%CI 1.90–3.90]; negative likelihood ratio was 0.19 [95% CI 0.14,-0.27]. The risk of publication bias was negligible with P  = 0.96. Conclusion Results of the meta-analysis suggest that mpMRI is a sensitive tool to diagnose prostate cancer. However, because of the high heterogeneity existing among the included studies, further studies are needed to apply the results of this meta-analysis in clinic.

Background The association between performance status (PS) and the prognosis of metastatic renal cell carcinoma (mRCC) patients receiving tyrosine kinase inhibitors (TKIs) remains controversial. The aim of this study is to evaluate the prognostic value of PS in mRCC patients treated with TKIs. Methods Electronic databases were searched to identify the studies that had assessed the association between pretreatment PS and prognosis in mRCC patients receiving TKIs. Hazard ratios (HRs) and 95% confidence interval (CI) for overall survival (OS) and progression-free survival (PFS) from eligible studies were used to calculate combined HRs. The heterogeneity across the included studies was assessed by Cochrane’s Q test and I 2 statistic. The Begg’s funnel plot and Egger’s linear regression teats were used to evaluate the potential publication bias. The meta-analysis was performed with RevMan 5.3 and Stata SE12.0 according to the PRISMA guidelines. Results A total of 6780 patients from 19 studies were included in this meta-analysis. The results showed that a poor PS was an effective prognostic factor of both OS (pooled HR: 2.08, 95% CI: 1.78–2.45) and PFS (pooled HR: 1.51, 95% CI: 1.20–1.91). Subgroup analysis revealed that poor PS significantly associated with poor OS and PFS in studies using Karnofsky PS scale (OS, pooled HR: 2.20, 95% CI: 1.65–2.94; PFS, pooled HR: 1.74, 95% CI: 1.19–2.56), conducted in Asia (OS, pooled HR: 2.25, 95% CI: 1.71–2.95; PFS, pooled HR: 1.73, 95% CI: 1.14–2.64) and Newcastle-Ottawa Scale score of 8 (OS, pooled HR: 2.61, 95% CI: 1.92–3.55; PFS, pooled HR: 2.43, 95% CI: 1.36–4.33). Conclusions This study suggests that a poor PS is significantly associated with poor prognosis in mRCC patients receiving TKIs.

Universal Design for Learning (UDL) has the potential to provide equal learning opportunities to various kinds of learners, which not only help address differential treatment in segregated education but also provide students with the same quality education. This study compares the current problems of integrated education in China in order to explore the necessity and countermeasures for the application of UDL in integrated education in China.

Purpose The purpose of this study was to characterize alterations in mucosa-associated microbiota in different anatomical locations of the stomach during gastric cancer progression and to identify associations between Helicobacter pylori infection and gastric microbial changes in patients with gastric cancer. Methods Twenty-five H. pylori negative subjects with chronic gastritis and thirty-four subjects with gastric cancer were recruited, including H. pylori negative and positive patients with tumors in the antrum and the corpus. Gastric mucosa-associated microbiota were determined by 16S ribosomal RNA gene sequencing using a 454 sequencing platform. Results We found that individuals with chronic gastritis from three different anatomical sites exhibited different microbiota compositions, although the microbial alpha diversity, richness and beta diversity were similar. Compared to patients with chronic gastritis, the gastric microbiota compositions were significantly different at the order level in the antrum and the corpus of patients with gastric cancer, which was dependent on the H. pylori infection status. Microbial alpha diversity and species richness, however, were similar between chronic gastritis and gastric cancer cases and independent of H. pylori status. The microbial community structure in patients with gastric cancer was distinct from that in patients with chronic gastritis. In addition, we found that the presence of H. pylori markedly altered the structure in gastric corpus cancer, but only mildly affected the antrum. Conclusion Our data revealed distinct niche-specific microbiota alterations during the progression from gastritis to gastric cancer. These alterations may reflect adaptions of the microbiota to the diverse specific environmental habitats in the stomach, and may play an important, as yet undetermined, role in gastric carcinogenesis.

Generalist embodied agents must perform interactive, causally-dependent reasoning, continually interacting with the environment, acquiring information, and updating plans to solve long-horizon tasks before they could be adopted in real-life scenarios. For instance, retrieving an apple from a cabinet may require opening multiple doors and drawers before the apple becomes visible and reachable, demanding sequential interaction under partial observability. However, existing benchmarks fail to systematically evaluate this essential capability. We introduce COIN, a benchmark designed to assess interactive reasoning in realistic robotic manipulation through three key contributions. First, we construct COIN-50: 50 interactive tasks in daily scenarios, and create COIN-Primitive required by causally-dependent tasks, and COIN-Composition with mid-term complexity for skill learning and generalization evaluation. Second, we develop a low-cost mobile AR teleoperation system and collect the COIN-Primitive Dataset with 50 demonstrations per primitive task (1,000 in total). Third, we develop systematic evaluation metrics about execution stability and generalization robustness to evaluate CodeAsPolicy, VLA, and language-conditioned H-VLA approaches. Our comprehensive evaluation reveals critical limitations in current methods: models struggle with interactive reasoning tasks due to significant gaps between visual understanding and motor execution. We provide fine-grained analysis of these limitations.

A formula of grain growth rate, based on a nonlinear capillarity-driven relation, is derived to predict and interpret realistic growth processes in polycrystalline systems. The derived formula reveals how the growth and stagnation of grains dominated by the correlated parameters (temperature, interfacial energy, step free energy, grain size and size distribution in polycrystalline system etc.). Our study provide a conclusive model of the growth and stagnation of grains, and thus offers helpful guides for the microstructural design to optimize the properties of polycrystalline materials.

Recent years have witnessed many successful trials in the robot learning field. For contact-rich robotic tasks, it is challenging to learn coordinated motor skills by reinforcement learning. Imitation learning solves this problem by using a mimic reward to encourage the robot to track a given reference trajectory. However, imitation learning is not so efficient and may constrain the learned motion. In this paper, we propose instruction learning, which is inspired by the human learning process and is highly efficient, flexible, and versatile for robot motion learning. Instead of using a reference signal in the reward, instruction learning applies a reference signal directly as a feedforward action, and it is combined with a feedback action learned by reinforcement learning to control the robot. Besides, we propose the action bounding technique and remove the mimic reward, which is shown to be crucial for efficient and flexible learning. We compare the performance of instruction learning with imitation learning, indicating that instruction learning can greatly speed up the training process and guarantee learning the desired motion correctly. The effectiveness of instruction learning is validated through a bunch of motion learning examples for a biped robot and a quadruped robot, where skills can be learned typically within several million steps. Besides, we also conduct sim-to-real transfer and online learning experiments on a real quadruped robot. Instruction learning has shown great merits and potential, making it a promising alternative for imitation learning.

Large Multimodal Models (LMMs) encode rich factual knowledge via cross-modal pre-training, yet their static representations struggle to maintain an accurate understanding of time-sensitive factual knowledge. Existing benchmarks remain constrained by static designs, inadequately evaluating LMMs' ability to understand time-sensitive knowledge. To address this gap, we propose MINED, a comprehensive benchmark that evaluates temporal awareness along 6 key dimensions and 11 challenging tasks: cognition, awareness, trustworthiness, understanding, reasoning, and robustness. MINED is constructed from Wikipedia by two professional annotators, containing 2,104 time-sensitive knowledge samples spanning six knowledge types. Evaluating 15 widely used LMMs on MINED shows that Gemini-2.5-Pro achieves the highest average CEM score of 63.07, while most open-source LMMs still lack time understanding ability. Meanwhile, LMMs perform best on organization knowledge, whereas their performance is weakest on sport. To address these challenges, we investigate the feasibility of updating time-sensitive knowledge in LMMs through knowledge editing methods and observe that LMMs can effectively update knowledge via knowledge editing methods in single editing scenarios.

Grain growth is a ubiquitous and fundamental phenomenon observed in the cellular structures with the grain assembly separated by a network of grain boundaries, including metals and ceramics. However, the underlying mechanism of grain growth has remained ambiguous for more than 60 years. The models for grain growth, based on the classically linear relationship between the grain boundary migration and capillary driving force, generally predict normal grain growth. Quantitative model for abnormal grain growth is lacking despite decades of efforts. Here, we present a unified model to reveal quantitatively how grain growth evolves, which predicts the normal, abnormal and stagnant behaviors of grain growth in polycrystalline materials. Our model indicates that the relationship between grain boundary migration and capillary driving force is generally nonlinear, but will switch to be the classically linear relationship in a specific case. Furthermore, the grain growth experiments observed in polycrystalline SrTiO3 demonstrates the validity of the unified model. Our study provides a unified, quantitative model to understand and predict grain growth in polycrystalline materials, and thus offers helpful guides for the microstructural design to optimize the properties of polycrystalline materials.