Development of a Large-Scale Roadside Facility Detection Model Based on the Mapillary Dataset.

The detection of road facilities or roadside structures is essential for high-definition (HD) maps and intelligent transportation systems (ITSs). With the rapid development of deep-learning algorithms in recent years, deep-learning-based object detection techniques have provided more accurate and efficient performance, and have become an essential tool for HD map reconstruction and advanced driver-assistance systems (ADASs). Therefore, the performance evaluation and comparison of the latest deep-learning algorithms in this field is indispensable. However, most existing works in this area limit their focus to the detection of individual targets, such as vehicles or pedestrians and traffic signs, from driving view images. In this study, we present a systematic comparison of three recent algorithms for large-scale multi-class road facility detection, namely Mask R-CNN, YOLOx, and YOLOv7, on the Mapillary dataset. The experimental results are evaluated according to the recall, precision, mean F1-score and computational consumption. YOLOv7 outperforms the other two networks in road facility detection, with a precision and recall of 87.57% and 72.60%, respectively. Furthermore, we test the model performance on our custom dataset obtained from the Japanese road environment. The results demonstrate that models trained on the Mapillary dataset exhibit sufficient generalization ability. The comparison presented in this study aids in understanding the strengths and limitations of the latest networks in multiclass object detection on large-scale street-level datasets.

Tegaserod maleate exhibits antileukemic activity by targeting TRPM8.

To identify therapeutic targets in acute myeloid leukemia (AML), we conducted growth inhibition screens of 2040 small molecules from a library of FDA-approved drugs using a panel of 12 AML cell lines. Tegaserod maleate, a 5-hydroxytryptamine 4 receptor partial agonist, elicits strong anti-AML effects in vitro and in vivo by targeting transient receptor potential melastatin subtype 8 (TRPM8), which plays critical roles in several important processes. However, the role of TRPM8 remains incompletely described in AML, whose treatment is based mostly on antimitotic chemotherapy. Here, we report an unexpected role of TRPM8 in leukemogenesis. Strikingly, TRPM8 knockout inhibits AML cell survival/proliferation by promoting apoptosis. Mechanistically, TRPM8 exerts its oncogenic effect by regulating the ERK-CREB/c-Fos signaling axis. Hyperactivation of ERK signaling can be reversed by TRPM8 inhibition. Importantly, TRPM8 is overexpressed in AML patients, indicating that it is a new prognostic factor in AML. Collectively, our work demonstrates the anti-AML effects of tegaserod maleate via targeting TRPM8 and indicates that TRPM8 is a regulator of leukemogenesis with therapeutic potential in AML.

Phase-resolved electrical detection of hybrid magnonic devices.

We demonstrate the electrical detection of magnon-magnon hybrid dynamics in yttrium iron garnet/permalloy (YIG/Py) thin film bilayer devices. Direct microwave current injection through the conductive Py layer excites the hybrid dynamics consisting of the uniform mode of Py and the first standing spin wave (n = 1) mode of YIG, which are coupled via interfacial exchange. Both the two hybrid modes, with Py or YIG dominated excitations, can be detected via the spin rectification signals from the conductive Py layer, providing phase resolution of the coupled dynamics. The phase characterization is also applied to a nonlocally excited Py device, revealing the additional phase shift due to the perpendicular Oersted field. Our results provide a device platform for exploring hybrid magnonic dynamics and probing their phases, which are crucial for implementing coherent information processing with magnon excitations.

Temporal variation of airborne fungi in university library rooms and its relation to environmental parameters and potential confounders.

Indoor airborne fungi have been associated with adverse human health effects. Therefore, it is important to understand the causes of underlying variation in airborne fungi in indoor environments. This study consequently aimed to investigate the association between indoor fungi with temporal variation, environmental parameters, and potential confounders over 10 months in four library rooms using Andersen samplers. Indoor fungal concentrations peaked in October and were lowest in March in both stack rooms, whereas the highest concentrations in both reading rooms were observed in September with lowest concentrations in July. Nonparametric analyses revealed higher fungal concentrations in the rooms that were significantly associated with relative humidity ≥ 60%, PM2.5 ≥ 35 µg/m3, number of people ≥ 16, open windows, working air conditioners, and room area < 400 m2. Multiple linear regression modeling for the library building considering only continuous variables revealed that relative humidity, PM2.5, and the number of people were significant predictors of fungal concentrations. Additionally, the model with continuous and categorical variables suggested that relative humidity, PM2.5, the number of people, ceiling fan condition, window state, and air conditioner operating status were significant predictor variables of concentrations. Outdoor fungal concentrations were a significant predictor for the two models of indoor fungal concentrations for each room. Ceiling fan or air conditioner operation was associated with altered fungal particle concentrations. These results provide a deeper understanding of indoor air fungal quality.

Hierarchical SrTiO3/NiFe2O4 composite nanostructures with excellent light response and magnetic performance synthesized toward enhanced photocatalytic activity.

Being capable of gathering advanced optical, electrical and magnetic properties originating from different components, multifunctional composite nanomaterials have been of concern increasingly. Herein, we have successfully demonstrated the preparation of SrTiO3/NiFe2O4 porous nanotubes (PNTs) and SrTiO3/NiFe2O4 particle-in-tubes (PITs) via a single-spinneret electrospinning and a side-by-side-spinneret electrospinning, respectively. The products were characterized by using scanning electron microscopy, transmission electron microscopy, X-ray diffraction, UV-visible diffuse reflectance spectra and a vibrating sample magnetometer in detail. The results indicate that SrTiO3/NiFe2O4 PNTs are the heterojunction nanotubes by connecting perovskite SrTiO3 and spinel NiFe2O4 nanoparticles, but SrTiO3/NiFe2O4 PITs are the self-assembled core/shell structures by embedding SrTiO3 nanoparticles into NiFe2O4 nanotubes. Compared with pure SrTiO3 nanofibers, the two SrTiO3/NiFe2O4 composites exhibit a powerful light response and excellent room temperature ferromagnetism. The magnetic separations directly reveal that such amazing recycling efficiencies of about 95% for SrTiO3/NiFe2O4 PNTs and about 99.5% for SrTiO3/NiFe2O4 PITs are obtained. Furthermore, both the magnetic composites perform considerable photocatalytic activity in the degradation of rhodamine B. We propose that Kirkendall-diffusion and phase-separation are probably responsible for the formation of SrTiO3/NiFe2O4 PITs, and this work could provide a feasible way to assemble the core/shell structures of different materials.