Physically Based Rendering of Synthetic Objects in Real Environments

Kronander, Joel

doi:10.3384/diss.diva-122588

Cited by 2 publications

(3 citation statements)

References 207 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One example of the most accurate approximations is the family of stochastic Monte Carlo methods. These are numerical methods based on averaging many random samples to calculate estimates of the true solution [136,185]. Another example of approximating the solution for light transport in a scene is the class of deterministic methods.…”

Section: Light Transport Simulationmentioning

confidence: 99%

Synthetic data for visual machine learning : A data-centric approach

Tsirikoglou

2022

Linköping Studies in Science and Technology. Dissertations

View full text Add to dashboard Cite

Deep learning allows computers to learn from observations, or else training data. Successful application development requires skills in neural network design, adequate computational resources, and a training data distribution that covers the application domain. We are currently witnessing an artificial intelligence (AI) outbreak with enough computational power to train very deep networks and build models that achieve similar or better than human performance. The crucial factor for the algorithms to succeed has proven to be the training data fed to the learning process. Too little or low quality or out-of-the-target distribution data will lead to poorly performing models no matter the capacity and the data regularization methods.This thesis is a data-centric approach to AI and presents a set of contributions related to synthesizing images for training supervised visual machine learning. It is motivated by the profound potential of synthetic data in cases of low availability of captured data, expensive acquisition and annotation, and privacy and ethical issues. The presented work aims to generate images similar to samples drawn from the target distribution and evaluate the generated data as the sole training data source and in conjunction with captured imagery. For this, two synthesis methods are explored: computer graphics and generative modeling. Computer graphics-based generation methods and synthetic datasets for computer vision tasks are thoroughly reviewed. In the same context, a system employing procedural modeling and physically-based rendering is introduced for data generation for urban scene understanding. The scheme is flexible, easily scalable, and produces complex and diverse images with pixel-perfect annotations at no cost. Generative Adversarial Networks (GANs) are also used to generate images for small data scenarios augmentation. The strategy advances the model's performance and robustness. Finally, ensembles of independently trained GANs investigate ways to improve images' diversity and create synthetic data to serve as the only training source.The application areas of the presented contributions relate to two image modalities, natural and histopathology images, to cover different aspects in the generation methods and the tasks' characteristics and requirements. There are showcased synthesized examples of natural images for automotive applications and weather classification, and histopathology images for breast cancer and colon adenocarcinoma metastasis detection. This thesis, as a whole, promotes data-centric supervised deep learning development by highlighting the potential of synthetic data as a training data resource. It emphasizes the control over the formation process, the ability of multi-modality formats, and the automatic generation of annotations.

show abstract

Section: Light Transport Simulationmentioning

confidence: 99%

Synthetic data for visual machine learning : A data-centric approach

Tsirikoglou

2022

Linköping Studies in Science and Technology. Dissertations

View full text Add to dashboard Cite

show abstract

“…This has led to the development of numerous VA techniques. Most of them combine geometrical acoustic (GA) simulation methods [1], like the image source method (ISM) [2] or the ray tracing method [3] in a hybrid model, where early and late reflections are modeled separately, depending on the requirements of the application and the desired accuracy and performance [4,5,6,7,8,9]. Other methods incorporate offline simulations which are executed prior to run-time for a grid of spatial points and then spatial interpolation is applied for the in-between positions [10,11].…”

Section: Introductionmentioning

confidence: 99%

“…Recently, the ISM has been implemented for interactive real-time VA applications with the help of high-performance computing. Optimizations for parallel CPU processing have been developed by Schroeder [5] and by Alarcão et al [16]. GPU implementation have also been considered.…”

Section: Introductionmentioning

confidence: 99%