Search for: All records

Underwater image enhancement and turbidity removal (dehazing) is a very challenging problem, not only due to the sheer variety of environments where it is applicable, but also due to the lack of high-resolution, labelled image data. In this paper, we present a novel, two-step deep learning approach for underwater image dehazing and colour correction. In iDehaze, we leverage computer graphics to physically model light propagation in underwater conditions. Specifically, we construct a three-dimensional, photorealistic simulation of underwater environments, and use them to gather a large supervised training dataset. We then train a deep convolutional neural network to remove the haze in these images, then train a second network to transform the colour space of the dehazed images onto a target domain. Experiments demonstrate that our two-step iDehaze method is substantially more effective at producing high-quality underwater images, achieving state-of-the-art performance on multiple datasets. Code, data and benchmarks will be open sourced. 

Physics-based via models are used in combination with neural network (NN) prediction of stripline to accurately estimate the full-channel response of trace-via structures for multilayered boards. Vias are modeled using analytical expressions for via capacitance and parallel plate impedance. Stripline response is predicted using a pretrained NN and converted into a circuit representation utilizing modal decomposition. The via-stripline transition is modeled with lumped elements. Individual elements are then concatenated using a segmentation approach. It is shown that such an approach can estimate frequency response up to 60 GHz. Correlation with full-wave solvers is demonstrated, and substantially improved computational speed is presented. 

This paper presents an enhanced closed-form approach for modeling and optimizing high-frequency PCB vias, implemented in Python and validated against industrystandard tools such as ADS and HFSS. The model incorporates resistance alongside inductance and capacitance to capture frequency-dependent losses and integrates non-functional pads (NFPs), demonstrating significant improvements in signal integrity by reducing reflections and enhancing return loss, particularly at 100 GHz. The methodology extends the frequency range of previous models from 100 GHz to 150 GHz, ensuring compatibility with next-generation standards like PCIe Gen 6. Validation results show insertion loss deviations under 3 dB and consistent return loss across the frequency range. The Python-based implementation offers a scalable and efficient solution for multilayer via designs, significantly reducing computational time compared to HFSS. This work provides a robust framework for high-speed PCB via modeling, with applications in academic research and industry, and includes future extensions to model differential signal vias. 

Transparent objects are a very challenging problem in computer vision. They are hard to segment or classify due to their lack of precise boundaries, and there is limited data available for training deep neural networks. As such, current solutions for this problem employ rigid synthetic datasets, which lack flexibility and lead to severe performance degradation when deployed on real-world scenarios. In particular, these synthetic datasets omit features such as refraction, dispersion and caustics due to limitations in the rendering pipeline. To address this issue, we present SuperCaustics, a real-time, open-source simulation of transparent objects designed for deep learning applications. SuperCaustics features extensive modules for stochastic environment creation; uses hardware ray-tracing to support caustics, dispersion, and refraction; and enables generating massive datasets with multi-modal, pixel-perfect ground truth annotations. To validate our proposed system, we trained a deep neural network from scratch to segment transparent objects in difficult lighting scenarios. Our neural network achieved performance comparable to the state-of-the-art on a real-world dataset using only 10% of the training data and in a fraction of the training time. Further experiments show that a model trained with SuperCaustics can segment different types of caustics, even in images with multiple overlapping transparent objects. To the best of our knowledge, this is the first such result for a model trained on synthetic data. Both our open-source code and experimental data are freely available online. 

In this paper, we present OpenWaters, a real-time open-source underwater simulation kit for generating photorealistic underwater scenes. OpenWaters supports creation of massive amount of underwater images by emulating diverse real-world conditions. It allows for fine controls over every variable in a simulation instance, including geometry, rendering parameters like ray-traced water caustics, scattering, and ground-truth labels. Using underwater depth (distance between camera and object) estimation as the use-case, we showcase and validate the capabilities of OpenWaters to model underwater scenes that are used to train a deep neural network for depth estimation. Our experimental evaluation demonstrates depth estimation using synthetic underwater images with high accuracy, and feasibility of transfer-learning of features from synthetic to real-world images. 

The dielectric substrate of printed circuit boards (PCBs) generally consists of epoxy or polyurethane resin and fiberglass, which can be modeled as inhomogeneous dielectric layers (IDLs). IDLs are not homogeneous in the perpendicular direction of the layers’ interface, which means that they constitute an anisotropic material. Thus, the effective dielectric constant (Dk) for IDLs is direction dependent, i.e., mode dependent. In this study, the effective Dk values of IDLs are extracted for different modes using three commonly used full-wave simulation models: the transverse electric (TE), transverse magnetic (TM), and equivalent parallel-plate capacitor models. The results for the TE and TM modes showed a 10% discrepancy. The effective Dk of IDLs was efficiently estimated by the parallel-plate capacitor model based on the layers’ dielectric properties and thicknesses. Moreover, the analytical formulas of resonance frequency for a cylindrical cavity resonator filled with IDLs were derived to extract the effective Dk of the IDLs at different resonance modes, which could be used as a guideline of resonator methods for PCB material characterization. These results correlated with the results from the equivalent parallel capacitor model. Finally, the impact of IDL anisotropy on transmission line loss and time-domain reflectometry impedance was analyzed based on full-wave simulation. 

This paper presents a simplified physics-based model for deep-snapback transient voltage suppressor (TVS) devices with open base BJT configuration. The proposed model is able to capture quasistatic IV curve as well as the transient behavior of the TVS including the initial voltage peak and the device turn on delay. A comparison of simulation and measurement results show that the proposed model can predict the quasi-static and transient behavior of an ESD protection device.