Search

Image captioning has become a hot topic in artificial intelligence research and sits at the intersection of computer vision and natural language processing. Most recent imaging captioning models have adopted an “encoder + decoder” architecture, in which the encoder is employed generally to extract the visual feature, while the decoder generates the descriptive sentence word by word. However, the visual features need to be flattened into sequence form before being forwarded to the decoder, and this results in the loss of the 2D spatial position information of the image. This limitation is particularly pronounced in the Transformer architecture since it is inherently not position-aware. Therefore, in this paper, we propose a simple coordinate-based spatial position encoding method (CSPE) to remedy this deficiency. CSPE firstly creates the 2D position coordinates for each feature pixel, and then encodes them by row and by column separately via trainable or hard encoding, effectively strengthening the position representation of visual features and enriching the generated description sentences. In addition, in order to reduce the time cost, we also explore a diagonal-based spatial position encoding (DSPE) approach. Compared with CSPE, DSPE is slightly inferior in performance but has a faster calculation speed. Extensive experiments on the MS COCO 2014 dataset demonstrate that CSPE and DSPE can significantly enhance the spatial position representation of visual features. CSPE, in particular, demonstrates BLEU-4 and CIDEr metrics improved by 1.6% and 5.7%, respectively, compared with a baseline model without sequence-based position encoding, and also outperforms current sequence-based position encoding approaches by a significant margin. In addition, the robustness and plug-and-play ability of the proposed method are validated based on a medical captioning generation model.

The seaworthiness of a ship plays a critical role in ensuring the safety of life and property and the prevention of marine pollution. One meaning of seaworthiness is the fundamental ability to sail safely to a destination, as well as being fit and ready for cargo. The parameters of seaworthiness have been expanded with the introduction of new maritime regulations. The entry of the 2019 amendments to the IMSBC Code into force has brought significant changes to bulk carriers with cargo that may liquefy at sea. Using examples from history, as well as an understanding of the environmental conditions of the oceans that ships are subject to, the seaworthiness of bauxite carriers was examined critically. First, the concept and doctrine of seaworthiness were clarified considering the existing transportation technology. Then, the spirit of the 2019 amendments to the IMSBC Code was applied to the seaworthiness of a bauxite carrier, showing that it demonstrated dynamic instability. Beyond that, from a transportability performance perspective, an extensive discussion was provided concerning the major correlation related to seaworthiness. Finally, practical recommendations were given on keeping a ship seaworthy before and during the process of bauxite transportation.

Feature pyramid networks and attention mechanisms are the mainstream methods to improve the detection performance of many current models. However, when they are learned jointly, there is a lack of information association between multi-level features. Therefore, this paper proposes a feature pyramid of the multi-level local attention method, dubbed as MLA-Net (Feature Pyramid Network with Multi-Level Local Attention for Object Detection), which aims to establish a correlation mechanism for multi-level local information. First, the original multi-level features are deformed and rectified using the local pixel-rectification module, and global semantic enhancement is achieved through the multi-level spatial-attention module. After that, the original features are further fused through the residual connection to achieve the fusion of contextual features to enhance the feature representation. Extensive ablation experiments were conducted on the MS COCO (Microsoft Common Objects in Context) dataset, and the results demonstrate the effectiveness of the proposed method with a 0.5% enhancement. An improvement of 1.2% was obtained on the PASCAL VOC (Pattern Analysis Statistical Modelling and Computational Learning, Visual Object Classes) dataset, reaching 81.8%, thereby, indicating that the proposed method is robust and can compete with other advanced detection models.

Cemented waste rock backfill (CWRB), which is a mixture of tailings, waste rock, cement, and water, is subjected to combination actions in underground mining operations and has been widely used in deep resource mining. While the strength requirement and macroscopic deformation behaviors of CWRB have been well studied, the mesoscopic damage evolution mechanisms are still not well understood. In this work, a CWRB sample with a waste rock proportion of 30% was studied with a uniaxial compression test under tomographic monitoring, using a 450 kV industrial X-ray computed tomography (CT). Clear CT images, CT value analysis, crack identification, and extraction reveal that CWRB damage evolution is extremely inhomogeneous and affected by the waste rock size, shape, and distribution. Furthermore, the crack initiation, propagation, and coalescence behaviors are limited to the existing waste rocks. When deformation grows to a certain extent, the cracks demonstrate an interlocking phenomenon and their propagation paths are affected by the waste rocks, which may improve the ability to resist compressive deformation. Volumetric dilatancy caused by the damage and cracking behavior has closed a link with the meso-structural changes, which are controlled by the interactions between the waste rocks and the cemented tailing paste.

A series of experimental flow tests for artificial block-in-matrix-soils (bimsoils) samples with various slenderness ratios were performed to study the Non-Darcy groundwater flow characteristics. The variations of seepage velocity, permeability coefficient, critical sample height, and non-Darcy flow factor for samples against slenderness ratios were investigated. A servo-controlled flow testing system that was developed by the authors was adopted to conduct the flow test. Cylindrical bimsoil samples (50 mm diameter and various heights) with staggered rock block percentages (RBPs) (30, 40, 50, and 60%, by mass) were prepared by compaction tests to roughly insure the same porosity of the soil matrix. The testing results show that flow the distance has a strong influence on the flow characteristics of bimsoil, and the relationship between the permeability coefficient and slenderness ratio is proposed. In addition, the critical sample height to eliminate the slenderness effect was determined, and the relationship between the critical sample height and RBP was established. Moreover, the responses of non-Darcy flow were studied by using an index of non-Darcy βfactor, which reveals the internal mechanism of the effect of flow distance on groundwater flow characteristics. The research results can be useful to the prediction of flow piping disaster for geological body made up of bimsoils.