Publications

Accurate classification of fine-grained images remains a challenge in backbones based on convolutional operations or self-attention mechanisms. This study proposes novel dual-current neural networks (DCNN), which combine the advantages of convolutional operations and self-attention mechanisms to improve the accuracy of fine-grained image classification. The main novel design features for constructing a weakly supervised learning backbone model DCNN include (a) extracting heterogeneous data, (b) keeping the feature map resolution unchanged, (c) expanding the receptive field, and (d) fusing global representations and local features. Experimental results demonstrated that using DCNN as the backbone network for classifying certain fine-grained benchmark datasets achieved performance advantage improvements of 13.5--19.5% and 2.2--12.9%, respectively, compared to other advanced convolution or attention-based fine-grained backbones.

A novel fuzzy reasoning method called the SQI (symmetric quintuple implicational) methodis put forward, which is a generalization of the QIP (quintuple implication principle) method. First of all, the symmetric quintuple implicational principles are presented, which are distinct from the ones of the QIP method. Then unified optimal solutions of the SQI method are obtained for FMP (fuzzy modus ponens) and FMT (fuzzy modus tollens), meanwhile corresponding reversible properties are verified. Furthermore, focusing on the case of multiple rules, optimal solutions of the SQI method are achieved, which involves two general approaches, i.e., FITA (first-infer-then-aggregate) and FATI (first-aggregate-then-infer). Equivalence relation of continuity and interpolation is analyzed for both FITA and FATI under the environment of the SQI method. Finally, one computing example arising in the field of affective computing is given for the SQI method with FATI. It is found that the SQI method preserves the same properties as the QIP method.

The resilient food supply chain (RFSC) has been identified as an effective model for mitigating food supply chain (FSC) risks. However, there exist many barriers impacting the implementation of the RFSC. Further, previous studies seldom utilize integrated decision models for identifying and ranking the barriers to implementing RFSC within uncertain environments. Thus, the study establishes an interval-valued Fermatean fuzzy (IVFF) decision framework to identify and rank these barriers. The framework is classified into four stages. First, to model the interaction between preference information, we introduce the IVFF-prioritized weighted average (PWA) operator to collect this information. Then, an integrated IVFF-CRITIC method is proposed to calculate the barrier weights considering their inter-correlation relationships. Next, the IVFF-PWA operator and IVFF-CRITIC method are incorporated into the PROMETHEE-II model to rank the barrier levels of alternative participation in the FSC. Further, a case study about analyzing implementation barriers to RFSC is employed to test the effectiveness and practicality of the presented framework. The result shows that the participation food processing company (priority: 0.161) has the highest barrier level. The findings of this article may offer decision support to stakeholders for mitigating the barriers to implementing a resilient supply chain in the food industry.

Unmanned aerial vehicles (UAVs) have emerged as promising platforms for fast, energy-efficient, and cost-effective package delivery. Path planning in 3-D urban environments is critical to drone delivery. The paper proposes a novel tangent-based (3D-TG) method for UAV path planning in 3-D urban environments. When a drone encounters an obstacle, a tangent graph is constructed to generate three sub-paths from both sides and above to bypass an obstacle, one of which is selected according to sophistically designed heuristic rules. The selected sub-path would be constantly adjusted its direction via tangent graph to avoid obstacles until the path can extend to the goal without obstacle collision. To avoid moving obstacles, velocity obstacle is incorporated in the 3D-TG. The experimental results on synthetic and realistic scenarios illustrate that 3D-TG performs well under static, unknown and dynamic environments. More significantly, 3D-TG can also generate a collision-free path for a drone to navigate through simple mazes efficiently, within a reasonable time.