Search results

This paper aims to present two different methods to speed up a test used in the sanitary ware industry that requires to count the number of granules that remains in the commodity after flushing. The test requires that 2,500 granules are added to the lavatory and less than 125 remain.

The problem is approached using two deep learning computer vision (CV) models. The first model is a Vision Transformers (ViT) classification approach and the second one is a U-Net paired with a connected components algorithm. Both models are trained and evaluated using a proprietary data set of 3,518 labeled images, and performance is compared.

It was found that both algorithms are able to produce competitive solutions. The U-Net algorithm achieves accuracy levels above 94% and the ViT model reach accuracy levels above 97%. At this time, the U-Net algorithm is being piloted and the ViT pilot is at the planning stage.

To the best of the authors’ knowledge, this is the first approach using CV to solve the granules problem applying ViT. In addition, this work updates the U-Net-Connected components algorithm and compares the results of both algorithms.

Fashion merchandising hold a paramount position within the realm of retail marketing. Currently, the purpose of this article is that the assessment of display effectiveness predominantly relies on the subjective judgment of merchandisers due to the absence of an effective evaluation method. Although eye-tracking devices have found extensive used in tracking the gaze trajectory of subject, they exhibit limitations in terms of stability when applied to the evaluation of various scenes. This underscores the need for a dependable, user-friendly and objective assessment method.

To develop a cost-effective and convenient evaluation method, the authors introduced an image processing framework for the assessment of variations in the impact of store furnishings. An optimized visual saliency methodology that leverages a multiscale pyramid model, incorporating color, brightness and orientation features, to construct a visual saliency heatmap. Additionally, the authors have established two pivotal evaluation indices aimed at quantifying attention coverage and dispersion. Specifically, bottom features are extract from 9 distinct scale images which are down sampled from merchandising photographs. Subsequently, these extracted features are amalgamated to form a heatmap, serving as the focal point of the evaluation process. The authors have proposed evaluation indices dedicated to measuring visual focus and dispersion, facilitating a precise quantification of attention distribution within the observed scenes.

In comparison to conventional saliency algorithm, the optimization method yields more intuitive feedback regarding scene contrast. Moreover, the optimized approach results in a more concentrated focus within the central region of the visual field, a pattern in alignment with physiological research findings. The results affirm that the two defined indicators prove highly effective in discerning variations in visual attention across diverse brand store displays.

The study introduces an intelligent and cost-effective objective evaluate method founded upon visual saliency. This pioneering approach not only effectively discerns the efficacy of merchandising efforts but also holds the potential for extension to the assessment of fashion advertisements, home design and website aesthetics.

Fixed mode noise and random mode noise always exist in the image sensor, which affects the imaging quality of the image sensor. The charge diffusion and color mixing between pixels in the photoelectric conversion process belong to fixed mode noise. This study aims to improve the image sensor imaging quality by processing the fixed mode noise.

Through an iterative training of an ergoable long- and short-term memory recurrent neural network model, the authors obtain a neural network model able to compensate for image noise crosstalk. To overcome the lack of differences in the same color pixels on each template of the image sensor under flat-field light, the data before and after compensation were used as a new data set to further train the neural network iteratively.

The comparison of the images compensated by the two sets of neural network models shows that the gray value distribution is more concentrated and uniform. The middle and high frequency components in the spatial spectrum are all increased, indicating that the compensated image edges change faster and are more detailed (Hinton and Salakhutdinov, 2006; LeCun et al., 1998; Mohanty et al., 2016; Zang et al., 2023).

In this paper, the authors use the iterative learning color image pixel crosstalk compensation method to effectively alleviate the incomplete color mixing problem caused by the insufficient filter rate and the electric crosstalk problem caused by the lateral diffusion of the optical charge caused by the adjacent pixel potential trap.

Powder bed additive manufacturing has recently seen substantial growth, yet consistently producing high-quality parts remains challenging. Recoating streaking is a common anomaly that impairs print quality. Several data-driven models for automatically detecting this anomaly have been proposed, each with varying effectiveness. However, comprehensive comparisons among them are lacking. Additionally, these models are often tailored to specific data sets. This research addresses this gap by implementing and comparing these anomaly detection models for recoating streaking in a reproducible way. This study aims to offer a clearer, more objective evaluation of their performance, strengths and weaknesses. Furthermore, this study proposes an improvement to the Line Profiles detection model to broaden its applicability, and a novel preprocessing step was introduced to enhance the models’ performances.

All found anomaly detection models have been implemented along with several preprocessing steps. Additionally, a new universal benchmarking data set has been constructed. Finally, all implemented models have been evaluated on this benchmarking data set and the effect of the different preprocessing steps was studied.

This comparison shows that the improved Line Profiles model established it as the most efficient detection approach in this study’s benchmark data set. Furthermore, while most state-of-the-art neural networks perform very well off the shelf, this comparison shows that specialised detection models outperform all others with the correct preprocessing.

This comparison gives new insights into different recoater streaking (RCS) detection models, showcasing each one with its strengths and weaknesses. Furthermore, the improved Line Profiles model delivers compelling performance in detecting RCS.

The purpose of this study is to advance the computational modeling of liquid splashing dynamics, while balancing simulation accuracy and computational efficiency, a duality often compromised in high-fidelity fluid dynamics simulations.

This study introduces the fluid efficient graph neural network simulator (FEGNS), an innovative framework that integrates an adaptive filtering layer and aggregator fusion strategy within a graph neural network architecture. FEGNS is designed to directly learn from extensive liquid splash data sets, capturing the intricate dynamics and intrinsically complex interactions.

FEGNS achieves a remarkable 30.3% improvement in simulation accuracy over traditional methods, coupled with a 51.6% enhancement in computational speed. It exhibits robust generalization capabilities across diverse materials, enabling realistic simulations of droplet effects. Comparative analyses and empirical validations demonstrate FEGNS’s superior performance against existing benchmark models.

The originality of FEGNS lies in its adaptive filtering layer, which independently adjusts filtering weights per node, and a novel aggregator fusion strategy that enriches the network’s expressive power by combining multiple aggregation functions. To facilitate further research and practical deployment, the FEGNS model has been made accessible on GitHub (https://github.com/nanjinyao/FEGNS/tree/main).

This paper aims to propose a stable and precise recognition and pose estimation method to deal with the difficulties that industrial parts often present, such as incomplete point cloud data due to surface reflections, lack of color texture features and limited availability of effective three-dimensional geometric information. These challenges lead to less-than-ideal performance of existing object recognition and pose estimation methods based on two-dimensional images or three-dimensional point cloud features.

In this paper, an image-guided depth map completion method is proposed to improve the algorithm's adaptability to noise and incomplete point cloud scenes. Furthermore, this paper also proposes a pose estimation method based on contour feature matching.

Through experimental testing on real-world and virtual scene dataset, it has been verified that the image-guided depth map completion method exhibits higher accuracy in estimating depth values for depth map hole pixels. The pose estimation method proposed in this paper was applied to conduct pose estimation experiments on various parts. The average recognition accuracy in real-world scenes was 88.17%, whereas in virtual scenes, the average recognition accuracy reached 95%.

The proposed recognition and pose estimation method can stably and precisely deal with the difficulties that industrial parts present and improve the algorithm's adaptability to noise and incomplete point cloud scenes.

The defect detection problem of color-patterned fabric is still a huge challenge due to the lack of manual defect labeling samples. Recently, many fabric defect detection algorithms based on feature engineering and deep learning have been proposed, but these methods have overdetection or miss-detection problems because they cannot adapt to the complex patterns of color-patterned fabrics. The purpose of this paper is to propose a defect detection framework based on unsupervised adversarial learning for image reconstruction to solve the above problems.

The proposed framework consists of three parts: a generator, a discriminator and an image postprocessing module. The generator is able to extract the features of the image and then reconstruct the image. The discriminator can supervise the generator to repair defects in the samples to improve the quality of image reconstruction. The multidifference image postprocessing module is used to obtain the final detection results of color-patterned fabric defects.

The proposed framework is compared with state-of-the-art methods on the public dataset YDFID-1(Yarn-Dyed Fabric Image Dataset-version1). The proposed framework is also validated on several classes in the MvTec AD dataset. The experimental results of various patterns/classes on YDFID-1 and MvTecAD demonstrate the effectiveness and superiority of this method in fabric defect detection.

It provides an automatic defect detection solution that is convenient for engineering applications for the inspection process of the color-patterned fabric manufacturing industry. A public dataset is provided for academia.

This study aims to describe the design and synthesis of two novel azo and imine chromophores-based dyes derived from two different aldehydes with intramolecular colour matching that are pH sensitive.

The visible absorption wavelength (λ_max) was extended when azo chromophore was included in imine-based systems. The dyed patterns created sophisticated colour-changing paper packaging sensors with pH-sensitive chromophores using alum as a mediator or mordant. Due to the tight adhesive bonding, the dyes on paper’s cellulose fibres could not be removed by ordinary water even at extremely high or low pH, which was confirmed by scanning electron microscopy analysis. The dyed patterns demonstrated an evident, sensitive and fast colour-changing mechanism with varying pH, from pale yellow to red for Dye-I and from pale yellow to brown-violet for Dye-II.

The λ_max for colour changing was recorded from 400 to 490 nm for Dye-I, whereas from 400 to 520 for Dye-II. The freshness judgement of food was checked using actual experiments with cooked crab spoilage, where the cooked crab was incubated at 37 ^oC for 6 h to see the noticeable colour change from yellow to brown-violet with Dye-II. The colour-changing mechanism was studied with Fourier transform infrared (FTIR) spectra at different pH, and thin layer chromatography, nuclear magnetic resonance and FTIR spectroscopy studied the desired structure formation of the dyes. Potential uses for smart packaging sensors include quickly detecting food freshness during transportation or right before consumption.

1. Two novel azo-imine dyes have been synthesized with a pH-responsive effect. 2. The pH-responsive mechanism was studied. 3. The study was supported by computational chemistry using density functional theory. 4. The obtained dyes were used to make pH-responsive sensors for seafood packaging to judge the freshness.

This study aims to address the global food waste problem in restaurants by analyzing customer sales information provided by restaurant tickets to gain valuable insights into directing sales of perishable products and optimizing product purchases according to customer demand.

A system based on unsupervised machine learning (ML) data models was created to provide a simple and interpretable management tool. This system performs analysis based on two elements: first, it consolidates and visualizes mutual and nontrivial relationships between information features extracted from tickets using multicomponent analysis, bootstrap resampling and ML domain description. Second, it presents statistically relevant relationships in color-coded tables that provide food waste-related recommendations to restaurant managers.

The study identified relationships between products and customer sales in specific months. Other ticket elements have been related, such as products with days, hours or functional areas and products with products (cross-selling). Big data (BD) technology helped analyze restaurant tickets and obtain information on product sales behavior.

This study addresses food waste in restaurants using BD and unsupervised ML models. Despite limitations in ticket information and lack of product detail, it opens up research opportunities in relationship analysis, cross-selling, productivity and deep learning applications.

The value and originality of this work lie in the application of BD and unsupervised ML technologies to analyze restaurant tickets and obtain information on product sales behavior. Better sales projection can adjust product purchases to customer demand, reducing food waste and optimizing profits.

This study aims to introduce a novel noise filter module designed for LiDAR simultaneous localization and mapping (SLAM) systems. The primary objective is to enhance pose estimation accuracy and improve the overall system performance in outdoor environments.

Distinct from traditional approaches, MCFilter emphasizes enhancing point cloud data quality at the pixel level. This framework hinges on two primary elements. First, the D-Tracker, a tracking algorithm, is grounded on multiresolution three-dimensional (3D) descriptors and adeptly maintains a balance between precision and efficiency. Second, the R-Filter introduces a pixel-level attribute named motion-correlation, which effectively identifies and removes dynamic points. Furthermore, designed as a modular component, MCFilter ensures seamless integration into existing LiDAR SLAM systems.

Based on rigorous testing with public data sets and real-world conditions, the MCFilter reported an increase in average accuracy of 12.39% and reduced processing time by 24.18%. These outcomes emphasize the method’s effectiveness in refining the performance of current LiDAR SLAM systems.

In this study, the authors present a novel 3D descriptor tracker designed for consistent feature point matching across successive frames. The authors also propose an innovative attribute to detect and eliminate noise points. Experimental results demonstrate that integrating this method into existing LiDAR SLAM systems yields state-of-the-art performance.