Power line inspection ensures the safe operation of a power transmission grid. Using unmanned aerial vehicle (UAV) images of power line corridors is an effective way to carry out these vital inspections. In this paper, we propose an automatic inspection method for power lines using UAV images. This method, known as the power line automatic measurement method based on epipolar constraints (PLAMEC), acquires the spatial position of the power lines. Then, the semi patch matching based on epipolar constraints (SPMEC) dense matching method is applied to automatically extract dense point clouds within the power line corridor. Obstacles can then be automatically detected by calculating the spatial distance between a power line and the point cloud representing the ground. Experimental results show that the PLAMEC automatically measures power lines effectively with a measurement accuracy consistent with that of manual stereo measurements. The height root mean square (RMS) error of the point cloud was 0.233 m, and the RMS error of the power line was 0.205 m. In addition, we verified the detected obstacles in the field and measured the distance between the canopy and power line using a laser range finder. The results show that the difference of these two distances was within ±0.5 m.

No abstract

Background: A relatively high early mortality rate (<30 days post-injury) for cervical spinal cord injury (SCI) has been observed. Objective: To investigate this early mortality rate observed after cervical SCI and analyze the associated influential factors. Methods: Medical records for 1163 patients with cervical SCI were reviewed, and the number of patients with early mortality was documented. Through logistic regression analysis, the effects of age, gender, occupation, cause of injury, severity of injury, highest involved spinal cord segment, nutritional condition during hospitalization, surgical treatment, tracheotomy, etc., on early mortality were assessed. Implementation of early treatment (i.e. surgery, tracheotomy, and nutritional support) and its effect on patient prognosis were also analyzed. Results: Early mortality occurred in 109 of 1163 patients (9.4%). Four factors affected the early mortality rate, including level and severity of SCI, whether or not surgery was performed, the time interval between SCI and surgery, malnutrition, and tracheotomy. Patients with an American Spinal Injury Association grade of A, a high cervical SCI (C1-C3), and/or no surgical intervention were statistically more likely to have early mortality (P < 0.001). Conclusion: Severe cervical SCI, upper-level cervical cord injury, malnutrition, and inappropriate tracheotomy are risk factors for early mortality in patients with cervical SCI. Surgery can reduce early mortality. Early tracheotomy should be performed in patients with complete upper-level cervical SCI, but patients with incomplete cervical SCI or complete low-level cervical SCI should initially be treated surgically to maintain smooth airway flow.

In existing visual representation learning tasks, deep convolutional neural networks (CNNs) are often trained on images annotated with single tag, such as ImageNet. However, single tag annotation cannot describe all important contents of one image, and some useful visual information may be wasted during training. In this work, we propose to train CNNs from images annotated with multiple tags, to enhance the quality of visual representation of the trained CNN model. To this end, we build a large-scale multi-label image database with 18M images and 11K categories, dubbed Tencent ML-Images. We efficiently train the ResNet-101 model with multi-label outputs on Tencent ML-Images, taking 90 hours for 60 epochs, based on a large-scale distributed deep learning framework, i.e., TFplus. The good quality of the visual representation of the Tencent ML-Images checkpoint is verified through three transfer learning tasks, including single-label image classification on ImageNet and Caltech-256, object detection on PASCAL VOC 2007, and semantic segmentation on PASCAL VOC 2012. The Tencent ML-Images database, the checkpoints of ResNet-101, and all the training code have been released at https://github.com/Tencent/tencent-ml-images. It is expected to promote other vision tasks in the research and industry community.

Most of the current antibody-drug conjugates (ADCs) in clinic are heterogeneous mixtures. To produce homogeneous ADCs, established procedures often require multiple steps or long reaction times. The introduced mutations or foreign sequences may cause high immunogenicity. Here, we explore a new concept of transforming CD38 enzymatic activity into a facile approach for generating site-specific ADCs. This was achieved through coupling bifunctional antibody-CD38 fusion proteins with designer dinucleotide-based covalent inhibitors with stably attached payloads. The resulting adenosine diphosphate–ribosyl cyclase–enabled ADC (ARC-ADC) with a drug-to-antibody ratio of 2 could be rapidly generated through single-step conjugation. The generated ARC-ADC targeting human epidermal growth factor receptor 2 (HER2) displays excellent stability and potency against HER2-positive breast cancer both in vitro and in vivo. This proof-of-concept study demonstrates a new strategy for production of site-specific ADCs. It may provide a general approach for the development of a novel class of ADCs with potentially enhanced properties.

No abstract

No abstract

The advantages of continuously and soundly obtaining large multidimensional, multiscale and multitemporal observation datasets from satellite remote sensing make it indispensable in building a national spatial data infrastructure. This paper introduces the ZY-3 satellite developed in China and discusses a fully automatic data-processing system to generate geoinformation products, such as digital elevation models (DEMs) and digital orthophotomaps (DOMs), based on ZY-3 imagery. The key technologies of automatic geoinformation product generation, including strip image-based bundle adjustment together with creating DEMs and DOMs, are illustrated. The accuracies of the georeferencing and automatically generated geoinformation products are also discussed. This automatic dataprocessing system is shown to provide a good foundation for near real-time derivation of such geoinformation products and for the promotion and application of Chinese domestic satellites.