No abstract

Feature selection (FS) is a classic and challenging optimization task in the field of machine learning and data mining. An equilibrium optimizer (EO) is a novel physics-based optimization algorithm; it was inspired by controlled volume mass balance models for estimating dynamic and equilibrium states. This paper presents two binary equilibrium optimizer algorithm and for selecting the optimal feature subset for classification problems. The first algorithm maps the continuous EO into a discrete type using S-shaped and V-shaped transfer functions (BEO-S and BEO-V). The second algorithm is based on the position of the current optimal solution (target) and position vector (BEO-T). To verify the performance of the proposed algorithm, 19 well-known UCI datasets are tested and compared with other advanced FS methods. The experimental results show that among the proposed binary EO algorithms, BEO-V2 has the best comprehensive performance and has better performance than other state-of-the-art metaheuristic algorithms in terms of the performance measures.

No abstract

In order to improve pesticide utilization rate and reduce the environmental pollution caused by pesticide ground loss, this paper proposes to use binocular vision to recognize the contour and distance information of fruit trees. To improve the recognition accuracy and speed, focusing on the optimization of SIFT stereo matching algorithm,A method for matching the feature points of left and right images base on cosine distance and the vector modulus is proposed. On this basis, two stereo matching algorithms are compared, The accuracy of the Improved SIFT stereo matching algorithm is improved by 1.53%, With this method, the recognition time is almost unchanged, And the stability of depth measurement is analyzed. When the target distance sensor is 180cm-220cm, the standard deviation is 1.3592cm, can meet the requirements of the work.

Summary Siamese‐based trackers have made great progress in visual tracking community, however, the shared structure of network between classification and regression tasks limits the ability of the trackers to obtain more robust classification prediction and more accurate regression prediction. In this paper, we propose an effective visual tracking framework (named Siamese Disentangled Tracking‐Head, SiamDTH), which disentangles classification and regression in Siamese‐based network for visual tracking from two aspects: feature decoupling and differentiated tracking‐head. First of all, we gather the features of receptive fields with different scales and ratios, and decouple the correlation features through two different styles of feature fusion mode for classification and regression respectively. Moreover, we design the differentiated tracking‐head structure in the sibling head for discriminately handling the parallel classification and regression tasks on visual tracking. Extensive experiments on visual tracking benchmarks including VOT2018, VOT2019 and OTB100 demonstrate that our proposed SiamDTH achieves state‐of‐the‐art performance with a considerable real‐time speed. Our source code is available at: https://github.com/xl0312/SiamDTH.