Autonomous underwater vehicles perceive environmental information relying on vision. However, the variety of the underwater environment always leads to image distortion. To effectively obtain color-corrected images that guarantee high sharpness and good content recovery, this paper proposes an algorithm based on deep learning networks. Our method is a modified version of generative adversarial networks for underwater image enhancement. Due to the complexity of the underwater environment, the green hue in the image is uneven, and the existing methods cannot recover the uneven green hue image well. Using our model, we generate corresponding depth estimation maps, along with the real underwater images, as inputs to the generator for color correction of monocular underwater images. The depth estimation maps, which are considered as image noise in the input, helps to distinguish the region of interest from the background. Furthermore, we formulate an objective function through several different loss functions, which can evaluate image quality globally. In addition, we performed training on the EUVP dataset, the comparison results show that our method is superior to other methods. Finally, we conduct extensive experiments on EUVP, UIEB and UFO-120 datasets, where our model not only significantly improves the quality of the underwater images, but also performs well over the state-of-the-art methods.

Anomaly detection remains a formidable challenge due to the inherent rarity and unpredictability of anomalous instances. Traditional knowledge distillation approaches, leveraging the comparison of output features from student-teacher (S-T) networks, often grapple with high false positive rates due to the phenomenon of overgeneralization. Addressing these limitations, our study introduces a novel approach, dubbed HKD-AD, with heterogeneity in both data and the structural composition of the S-T networks. Specifically, we fed paired normal and anomalous images to the S-T networks, with the student network aimed at mimicking normal teacher features from the simulated abnormal images. We also employ the Transform-based teacher to guide the CNN-based student network, which not only amplifies feature extraction capabilities but also substantially enhances anomaly localization. Experiments conducted on the challenging MVTec AD dataset demonstrate that our HKD-AD significantly outperforms existing methods, achieving state-of-the-art results with high efficiency.

In the traditional customized product (CP) configuration design system, the configuration rules in the database usually rely on the manual input and maintenance of experienced designers. In complex product design, configuration rules are numerous, complex, and difficult to understand. The way of human input based on experience consumes plenty of resources, which are strictly limited by experienced designers. This overreliance on those experienced designers has seriously restricted the development of enterprises. To address this problem, a least recently used dynamic decision tree (LRU-DDT) algorithm for CP configuration rule intelligent extraction and dynamic updating is put forward in this article. Based on the decision table majority (DTM) classifier, the condition attributes are first reduced. An improved J48 decision tree algorithm is proposed to extract the CP configuration rules. In addition, the CP configuration rules can be dynamically updated based on LRU-DDT. To verify the proposed method, the GB10 high-speed elevator configuration design process is taken as an example. The configuration rules are extracted and updated by the proposed algorithms. The results show that the configuration solution efficiency is improved by 16%.

Anomaly detection is inherently challenging, as anomalies typically emerge only at test time. While reconstruction-based methods are popular, their reliance on CNN backbones with local receptive fields limits discrimination and precise localization. We propose FFC-AD, a reconstruction framework using Fourier Feature Convolutions (FFCs) to capture global information early, and we introduce Hidden Space Anomaly Simulation (HSAS), a latent-space regularization strategy that mitigates overgeneralization. Experiments on MVTec AD and VisA demonstrate that FFC-AD significantly outperforms state-of-the-art methods in both detection and segmentation accuracy.

With the rapid advancement of intelligent manufacturing technologies, the operating environments of modern robotic arms are becoming increasingly complex. In addition to the diversity of objects, there is often a high degree of similarity between the foreground and the background. Although traditional RGB-based object-detection models have achieved remarkable success in many fields, they still face the challenge of effectively detecting targets with textures similar to the background. To address this issue, we introduce the WoodenCube dataset, which contains over 5000 images of 10 different types of blocks. All images are densely annotated with object-level categories, bounding boxes, and rotation angles. Additionally, a new evaluation metric, Cube-mAP, is proposed to more accurately assess the detection performance of cube-like objects. In addition, we have developed a simple, yet effective, framework for WoodenCube, termed CS-SKNet, which captures strong texture features in the scene by enlarging the network’s receptive field. The experimental results indicate that our CS-SKNet achieves the best performance on the WoodenCube dataset, as evaluated by the Cube-mAP metric. We further evaluate the CS-SKNet on the challenging DOTAv1.0 dataset, with the consistent enhancement demonstrating its strong generalization capability.