Search

A thorough understanding of media tightness in automotive electronics is crucial for ensuring more reliable and compact product designs, ultimately improving product quality. Concerning the fundamental characteristics of fluid leakage issues, the dynamic wetting and penetration behavior on small scales is of special interest and importance. In this work, four T-shaped microchannels with one inlet and two outlets are experimentally investigated in terms of contact angle dynamics and interface movement over time, generating novel insight into the wetting mechanisms and fluid distribution. With a main channel width of 1 mm, a crevice width of w = 0.3 mm, 0.4 mm and a rounding edge radius of r = 0.1 mm, 0.2 mm, the geometrical effects on the fluid penetration depth in the crevice and the interface edge pinning effect are analyzed quantitatively using an automated image processing procedure. It is found that the measured dynamic contact angles in all parts can be well described by molecular kinetic theory using local contact line velocities, even with local surface effects and abrupt geometry changes. Moreover, a smaller crevice width, a sharper edge and a larger flow velocity tend to enhance the interface pinning effect and prevent fluid penetration into the crevice. The rounding radius has a more significant effect on the interface pinning compared with crevice width. The experimental data and image processing algorithm are made publicly available.

The efficient and voidless filling of microcavities is of great importance for Lab-on-a-Chip applications. However, predicting whether microcavities will be filled or not under different circumstances is still difficult due to the local flow effects dominated by surface tension. In this work, a close-up study of the microcavity filling process is presented, shedding light on the mechanisms of the filling process using experimental insights accompanied by 3D numerical simulations. The movement of a fluid interface over a microcavity array is investigated optically under consideration of different fluids, capillary numbers, and cavity depths, revealing a regime map of different filling states. Moreover, the transient interface progression over the cavities is analyzed with attention to small-scale effects such as pinning. Besides the visual analysis of the image series, quantitative data of the dynamic contact angle and the interface progression is derived using an automated evaluation workflow. In addition to the experiments, 3D Volume-of-Fluid simulations are employed to further investigate the interface shape. It is shown that the simulations can not only predict the filling states in most cases, but also the transient movement and shape of the interface. The data and code associated with this work are publicly available at Bosch Research GitHub and at the TUDatalib data repository., Comment: submitted to International Journal of Multiphase Flow

Preventing fluid penetration poses a challenging reliability concern in the context of power electronics, which is usually caused by unforeseen microfractures along the sealing joints. A better and more reliable product design heavily depends on the understanding of the dynamic wetting processes happening inside these complex microfractures, i.e. microchannels. A novel automated image processing procedure is proposed in this work for analyzing the moving interface and the dynamic contact angle in microchannels. In particular, the developed method is advantageous for experiments involving non-transparent samples, where extracting the fluid interface geometry poses a significant challenge. The developed method is validated with theoretical values and manual measurements and exhibits high accuracy. The implementation is made publicly available. The developed method is validated and applied to experimental investigations of forced wetting with two working fluids (water and 50 wt% glycerin/water mixture) in four distinct microchannels characterized by different dimensions and curvature. The comparison between the experimental results and molecular kinetic theory (MKT) reveals that the dynamic wetting behavior can be described well by MKT, even in highly curved microchannels. The dynamic wetting behavior shows a strong dependency on the channel geometry and curvature.

Diffusion models, emerging as powerful deep generative tools, excel in various applications. They operate through a two-steps process: introducing noise into training samples and then employing a model to convert random noise into new samples (e.g., images). However, their remarkable generative performance is hindered by slow training and sampling. This is due to the necessity of tracking extensive forward and reverse diffusion trajectories, and employing a large model with numerous parameters across multiple timesteps (i.e., noise levels). To tackle these challenges, we present a multi-stage framework inspired by our empirical findings. These observations indicate the advantages of employing distinct parameters tailored to each timestep while retaining universal parameters shared across all time steps. Our approach involves segmenting the time interval into multiple stages where we employ custom multi-decoder U-net architecture that blends time-dependent models with a universally shared encoder. Our framework enables the efficient distribution of computational resources and mitigates inter-stage interference, which substantially improves training efficiency. Extensive numerical experiments affirm the effectiveness of our framework, showcasing significant training and sampling efficiency enhancements on three state-of-the-art diffusion models, including large-scale latent diffusion models. Furthermore, our ablation studies illustrate the impact of two important components in our framework: (i) a novel timestep clustering algorithm for stage division, and (ii) an innovative multi-decoder U-net architecture, seamlessly integrating universal and customized hyperparameters., Comment: The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2024

In this work, we investigate an intriguing and prevalent phenomenon of diffusion models which we term as "consistent model reproducibility": given the same starting noise input and a deterministic sampler, different diffusion models often yield remarkably similar outputs. We confirm this phenomenon through comprehensive experiments, implying that different diffusion models consistently reach the same data distribution and scoring function regardless of diffusion model frameworks, model architectures, or training procedures. More strikingly, our further investigation implies that diffusion models are learning distinct distributions affected by the training data size. This is supported by the fact that the model reproducibility manifests in two distinct training regimes: (i) "memorization regime", where the diffusion model overfits to the training data distribution, and (ii) "generalization regime", where the model learns the underlying data distribution. Our study also finds that this valuable property generalizes to many variants of diffusion models, including those for conditional use, solving inverse problems, and model fine-tuning. Finally, our work raises numerous intriguing theoretical questions for future investigation and highlights practical implications regarding training efficiency, model privacy, and the controlled generation of diffusion models., Comment: NeurIPS Diffusion Model Workshop 2023 (best paper award), the Forty-first International Conference on Machine Learning (ICML 2024)

Transparent objects present multiple distinct challenges to visual perception systems. First, their lack of distinguishing visual features makes transparent objects harder to detect and localize than opaque objects. Even humans find certain transparent surfaces with little specular reflection or refraction, like glass doors, difficult to perceive. A second challenge is that depth sensors typically used for opaque object perception cannot obtain accurate depth measurements on transparent surfaces due to their unique reflective properties. Stemming from these challenges, we observe that transparent object instances within the same category, such as cups, look more similar to each other than to ordinary opaque objects of that same category. Given this observation, the present paper explores the possibility of category-level transparent object pose estimation rather than instance-level pose estimation. We propose \textit{\textbf{TransNet}}, a two-stage pipeline that estimates category-level transparent object pose using localized depth completion and surface normal estimation. TransNet is evaluated in terms of pose estimation accuracy on a large-scale transparent object dataset and compared to a state-of-the-art category-level pose estimation approach. Results from this comparison demonstrate that TransNet achieves improved pose estimation accuracy on transparent objects. Moreover, we use TransNet to build an autonomous transparent object manipulation system for robotic pick-and-place and pouring tasks.

Parasitic currents still pose a significant challenge for the investigation of two-phase flow in Lab-on-Chip (LoC) applications with Volume-of-Fluid (VoF) simulations. To counter the impact of such spurious velocity fields in the vicinity of the fluid interface, this work presents an implementation of an artificial interface viscosity model in OpenFOAM. The model is introduced as an additional dampening term in the momentum conservation equation. It is implemented as a fvOption, allowing for its simple application to existing VoF solvers. Validation is performed with hydrodynamic and wetting cases, in which constant artificial viscosity values are prescribed to examine the sensitivity of the solution to the artificial dampening. The artificial viscosity model shows promising results in reducing spurious currents for two considered geometrical VoF solvers, namely interIsoFoam and InterFlow. It is found that the influence of the artificial viscosity heavily depends on the fluid properties. Applying the model to simulations of an interface traversing through microcavities relevant in LoC applications, experimental results of the interface progression are predicted well, while spurious currents are effectively reduced by approximately one order of magnitude due to the artificial viscosity model. The code is publicly available on GitHub (https://github.com/boschresearch/sepMultiphaseFoam/tree/publications/ArtificialInterfaceViscosity)., Comment: author manuscript of published article

Transparent objects present multiple distinct challenges to visual perception systems. First, their lack of distinguishing visual features makes transparent objects harder to detect and localize than opaque objects. Even humans find certain transparent surfaces with little specular reflection or refraction, e.g. glass doors, difficult to perceive. A second challenge is that common depth sensors typically used for opaque object perception cannot obtain accurate depth measurements on transparent objects due to their unique reflective properties. Stemming from these challenges, we observe that transparent object instances within the same category (e.g. cups) look more similar to each other than to ordinary opaque objects of that same category. Given this observation, the present paper sets out to explore the possibility of category-level transparent object pose estimation rather than instance-level pose estimation. We propose TransNet, a two-stage pipeline that learns to estimate category-level transparent object pose using localized depth completion and surface normal estimation. TransNet is evaluated in terms of pose estimation accuracy on a recent, large-scale transparent object dataset and compared to a state-of-the-art category-level pose estimation approach. Results from this comparison demonstrate that TransNet achieves improved pose estimation accuracy on transparent objects and key findings from the included ablation studies suggest future directions for performance improvements.

Transparent objects are ubiquitous in household settings and pose distinct challenges for visual sensing and perception systems. The optical properties of transparent objects leave conventional 3D sensors alone unreliable for object depth and pose estimation. These challenges are highlighted by the shortage of large-scale RGB-Depth datasets focusing on transparent objects in real-world settings. In this work, we contribute a large-scale real-world RGB-Depth transparent object dataset named ClearPose to serve as a benchmark dataset for segmentation, scene-level depth completion and object-centric pose estimation tasks. The ClearPose dataset contains over 350K labeled real-world RGB-Depth frames and 5M instance annotations covering 63 household objects. The dataset includes object categories commonly used in daily life under various lighting and occluding conditions as well as challenging test scenarios such as cases of occlusion by opaque or translucent objects, non-planar orientations, presence of liquids, etc. We benchmark several state-of-the-art depth completion and object pose estimation deep neural networks on ClearPose. The dataset and benchmarking source code is available at https://github.com/opipari/ClearPose., Comment: ECCV 2022 accepted paper

Visual perception tasks often require vast amounts of labelled data, including 3D poses and image space segmentation masks. The process of creating such training data sets can prove difficult or time-intensive to scale up to efficacy for general use. Consider the task of pose estimation for rigid objects. Deep neural network based approaches have shown good performance when trained on large, public datasets. However, adapting these networks for other novel objects, or fine-tuning existing models for different environments, requires significant time investment to generate newly labelled instances. Towards this end, we propose ProgressLabeller as a method for more efficiently generating large amounts of 6D pose training data from color images sequences for custom scenes in a scalable manner. ProgressLabeller is intended to also support transparent or translucent objects, for which the previous methods based on depth dense reconstruction will fail. We demonstrate the effectiveness of ProgressLabeller by rapidly create a dataset of over 1M samples with which we fine-tune a state-of-the-art pose estimation network in order to markedly improve the downstream robotic grasp success rates. ProgressLabeller is open-source at https://github.com/huijieZH/ProgressLabeller., Comment: IROS 2022 accepted paper; project page: https://progress.eecs.umich.edu/projects/progress-labeller/