Jürgen Döllner

Universität Potsdam · Chair for Computer Graphics

We present a methodology for conditional control of human shape and pose in pretrained text-to-image diffusion models using a 3D human parametric model (SMPL). Fine-tuning these diffusion models to adhere to new conditions requires large datasets and high-quality annotations, which can be more cost-effectively acquired through synthetic data genera...

Point clouds are widely used as a versatile representation of 3D entities and scenes for all scale domains and in a variety of application areas, serving as a fundamental data category to directly convey spatial features. However, due to point sparsity, lack of structure, irregular distribution, and acquisition-related inaccuracies, results of poin...

Deep learning models achieve high accuracy in the semantic segmentation of 3D point clouds; however, it is challenging to discern which patterns a model has learned and how it derives its output from the input. Recently, the Integrated Gradients method has been adopted to explain semantic segmentation models for 3D point clouds. This method can be...

This paper contributes an optimized web-based rendering approach and implementation for parameterized meshes used as 3D glyphs for information visualization. The approach is based on geometry instancing by means of instanced rendering in three.js, and further allows for dynamic mesh selection according to a level-of-detail function and data-driven...

This paper contributes an optimized web-based rendering approach and implementation for parameterized meshes used as 3D glyphs for information visualization. The approach is based on geometry instancing by means of instanced rendering in three.js, and further allows for dynamic mesh selection according to a level-of-detail function and data-driven...

Dimensionality reductions are a class of unsupervised learning algorithms that aim to find a lower-dimensional embedding for a high-dimensional dataset while preserving local and global structures. By representing a high-dimensional dataset as a two-dimensional scatterplot, a user can explore structures within the dataset. However, dimensionality r...

Ground point filtering on national-level datasets is a challenge due to the presence of multiple types of landscapes. This limitation does not simply affect to individual users, but it is in particular relevant for those national institutions in charge of providing national-level Light Detection and Ranging (LiDAR) point clouds. Each type of landsc...

The semantic similarity between documents of a text corpus can be visualized using map-like metaphors based on two-dimensional scatterplot layouts. These layouts result from a dimensionality reduction on the document-term matrix or a representation within a latent embedding, including topic models. Thereby, the resulting layout depends on the input...

While methods for generative image synthesis and example-based stylization produce impressive results, their black-box style representation intertwines shape, texture, and color aspects, limiting precise stylistic control and editing of artistic images. We introduce a novel method for decomposing the style of an artistic image that enables interact...

3D point clouds acquired with terrestrial or mobile LiDAR sensors are increasingly used to map urban forests. The segmentation of separate tree instances, i.e., subsets of points representing individual trees, is a relevant step in automatically extracting tree inventory data from 3D point clouds. Various algorithms have been proposed for tree inst...

Standard datasets are frequently used to train and evaluate Machine Learning models. However, the assumed standardness of these datasets leads to a lack of in-depth discussion on how their labels match the derived categories for the respective use case. In other words, the standardness of the datasets seems to fog coherency and applicability, thus...

In 2016, BitMEX introduced a novel type of crypto derivates – Perpetual Swaps, i.e., futures with an infinite term. Perpetual swaps provide a new strategic risk management tool for cryptocurrencies due to their custody-free nature, high leverage, and funding mechanism, but there has been little quantitative analysis on the their benefits. In this p...

Point clouds are widely used as a versatile representation of 3D entities and scenes for all scale domains and in a variety of application areas, serving as a fundamental data category to directly convey spatial features. However, due to point sparsity, lack of structure, irregular distribution, and acquisition-related inaccuracies, results of poin...

Machine Learning models underlie a trade-off between accurracy and explainability. Given a trained, complex model, we contribute a dashboard that supports the process to derive more explainable models, here: Fast-and-Frugal Trees, with further introspection using feature importances and spurious correlation analyses. The dashboard further allows to...

Machine Learning models underlie a trade-off between accurracy and explainability. Given a trained, complex model, we contribute a dashboard that supports the process to derive more explainable models, here: Fast-and-Frugal Trees, with further introspection using feature importances and spurious correlation analyses. The dashboard further allows to...

Dimensionality Reduction Techniques (DRs) are used for projecting high-dimensional data onto a two-dimensional plane. One subclass of DRs are such techniques that utilize landmarks. Landmarks are a subset of the original data space that are projected by a slow and more precise technique. The other data points are then placed in relation to these la...

Dimensionality Reduction Techniques (DRs) are used for projecting high-dimensional data onto a two-dimensional plane. One subclass of DRs are such techniques that utilize landmarks. Landmarks are a subset of the original data space that are projected by a slow and more precise technique. The other data points are then placed in relation to these la...

Most cryptocurrency spot trading occurs on centralized crypto exchanges, where offers for buying and selling are organized via an order book. In liquid markets, the price achieved for buying and selling deviates only slightly from the assumed reference price, i.e., trading is associated with low implicit costs. However, compared to traditional fina...

Programmers who want to explore the architecture of software systems need appropriate visualizations such as software maps. However, existing software visualizations mainly display the static software structure, neglecting important dynamic runtime behavior. We propose animated 2.5D object maps that depict particular objects and their interactions...

Text spatializations for text corpora often rely on two-dimensional scatter plots generated from topic models and dimensionality reductions. Topic models are unsupervised learning algorithms that identify clusters, so-called topics, within a corpus, representing the underlying concepts. Furthermore, topic models transform documents into vectors, ca...

3D point clouds are a widely used representation for surfaces and object geometries. However, their visualization can be challenging due to point sparsity and acquisition inaccuracies, leading to visual complexity and ambiguity. Non-photorealistic rendering (NPR) addresses these challenges by using stylization techniques to abstract from certain de...

Program comprehension is a key activity in software development. Several visualization approaches such as software maps have been proposed to support programmers in exploring the architecture of software systems. However, for the exploration of program behavior, programmers still rely on traditional code browsing and debugging tools to build a ment...

3D point clouds are a widely used representation for surfaces and object geometries. However, their visualization can be challenging due to point sparsity and acquisition inaccuracies, leading to visual complexity and ambiguity. Non-photorealistic rendering (NPR) addresses these challenges by using stylization techniques to abstract from certain de...

Program comprehension is a key activity in software development. Several visualization approaches such as software maps have been proposed to support programmers in exploring the architecture of software systems. However, for the exploration of program behavior, programmers still rely on traditional code browsing and debugging tools to build a ment...

Text spatializations for text corpora often rely on two-dimensional scatter plots generated from topic models and dimensionality reductions. Topic models are unsupervised learning algorithms that identify clusters, so-called topics, within a corpus, representing the underlying concepts. Furthermore, topic models transform documents into vectors, ca...

Digital twins that serve as virtual representations of real-world objects and structures, are used in various applications for urban environments. Challenges for creating and maintaining digital twins involve data acquisition, fusion of heterogeneous data types, AI-based data analysis, and the integration into existing applications and workflows. I...

Readily available software analysis and analytics tools are often operated within external services, where the measured software analysis data is kept internally and no external access to the data is available. We propose an approach to integrate visual software analysis on the GitHub platform by leveraging GitHub Actions and the GitHub API, coveri...

Video Moment Retrieval (VMR) is a challenging task at the intersection of vision and language, with the goal to retrieve relevant moments from videos corresponding to natural language queries. State-of-the-art approaches for VMR often rely on large amounts of training data including frame-level saliency annotations, weakly supervised pre-training o...

As machine learning models are becoming more widespread and see use in high-stake decisions, the explainability of these decisions is getting more relevant. One approach for explainability are counterfactual explanations, which are defined as changes to a data point such that it appears as a different class. Their close connection to the original d...

Text synthesis tools are becoming increasingly popular and better at mimicking human language. In trust-sensitive decisions, such as plagiarism and fraud detection, identifying AI-generated texts poses larger difficulties: decisions need to be made explainable to ensure trust and accountability. To support users in identifying AI-generated texts, w...

Topic models are a class of unsupervised learning algorithms for detecting the semantic structure within a text corpus. Together with a subsequent dimensionality reduction algorithm, topic models can be used for deriving spatializations for text corpora as two-dimensional scatter plots, reflecting semantic similarity between the documents and suppo...

This paper introduces an art-directable stroke-based rendering technique for transforming photos into painterly renditions on mobile devices. Unlike previous approaches that rely on time-consuming iterative computations and explicit brush-stroke geometry, our method offers an interactive image-based implementation tailored to the capabilities of mo...

Presentation of the research paper "Controlling Geometric Abstraction and Texture for Artistic Images" at the 22th International Conference on Cyberworlds (CW2023) in Sousse, Tunisia.

Building Information Modelling (BIM) is becoming increasingly prevalent in infrastructure asset management, as it facilitates current management practices. This includes the construction of BIM models for roads, rails, bridges, tunnels etc. Bridges are particularly challenging to digitalize due to their complex geometry. The manual construction of...

Digital twins that serve as virtual representations of real-world objects and structures, are used in various applications for urban environments. Challenges for creating and maintaining digital twins involve data acquisition, fusion of heterogeneous data types, AI-based data analysis, and the integration into existing applications and workflows. I...

We present a novel method for the interactive control of geometric abstraction and texture in artistic images. Previous example-based stylization methods often entangle shape, texture, and color, while generative methods for image synthesis generally either make assumptions about the input image, such as only allowing faces or do not offer precise...

Image stylization has seen significant advancement and widespread interest over the years, leading to the development of a multitude of techniques. Extending these stylization techniques, such as Neural Style Transfer (NST), to videos is often achieved by applying them on a per‐frame basis. However, per‐frame stylization usually lacks temporal cons...

Readily available software analytics tools are often operated as external services, where measured software analysis data is kept internal , and no external use of the data is available. We propose an approach for embedded visual software analytics on the GitHub platform by leaveraging GitHub Actions and the GitHub API. The proposed approach collec...

Ray tracing remains of interest to Computer Graphics community with its elegant framing of how light interacts with objects, being able to easily support multiple light sources, and simple framework of merging synthetic and real cameras. Recent trends to provide implementations at the chip-level means raytracing’s constant quest of realism would pr...

This paper presents a framework and mobile video editing app for interactive artistic augmentation of human motion in videos. While creating motion effects with industry-standard software is time-intensive and requires expertise, and popular video effect apps have limited customization options, our approach enables a multitude of art-directable, hi...

Presentation of the research paper "Interactive Control over Temporal Consistency while Stylizing Video Streams" at the 34th Eurographics Symposium on Rendering (EGSR 2023) in Delft, Netherlands.

This contribution presents the use of a newly developed approach for the generation and modelling of spatially referenced, multidimensional virtual city models from remote sensing data, exemplified by the case of East Jerusalem. The focus is on the effective application of the entire process from data acquisition, pre-processing and management to s...

In urban environments, roadside vegetation provides important ecosystem services. Reliable and up-to-date information on urban vegetation is therefore needed as a basis for sustainable urban design and regular tasks such as vegetation maintenance. Mobile laser scanning (MLS), i. e., the use of vehicle-mounted laser scanners, offers strong potential...

The Hilbert and Moore treemap layout algorithms are based on the space-filling Hilbert and Moore curves, respectively, to map tree-structured datasets to a 2D treemap layout. Considering multiple snapshots of a time-variant dataset, one of the design goals for Hilbert and Moore treemaps is layout stability, i.e., low changes in the layout for low c...

The Hilbert and Moore treemap layout algorithms are based on the space-filling Hilbert and Moore curves, respectively, to map tree-structured datasets to a 2D treemap layout. Considering multiple snapshots of a time-variant dataset, one of the design goals for Hilbert and Moore treemaps is layout stability, i.e., low changes in the layout for low c...

Quali-quantitative methods provide ways for interrogating Convolutional Neural Networks (CNN). For it, we propose a dashboard using a quali-quantitative method based on quantitative metrics and saliency maps. By those means, a user can discover patterns during the training of a CNN. With this, they can adapt the training hyperparameters of the mode...

LiDAR scanning technology is an established method for capturing landscapes, buildings, or roads in order to create a so-called spatial digital twin of the reality, stored as a large collection of 3D coordinates called 3D point cloud. This spatial data offers high density and precision at the cost of hard to extract shape or object information. One...

Image acquisition in low-light conditions suffers from poor quality and significant degradation in visual aesthetics. This affects the visual perception of the acquired image and the performance of computer vision and image processing algorithms applied after acquisition. Especially for videos, the additional temporal domain makes it more challengi...

Presentation of the research contribution "ALIVE: Adaptive Chromaticity for Interactive Low-light Image and Video Enhancement" at the 31. International Conference on Computer Graphics, Visualization and Computer Vision (WSCG 2023).

Presentation of the research contribution "A Framework for Art directed Augmentation of Human Motion in Videos on Mobile Devices" at the 31. International Conference on Computer Graphics, Visualization and Computer Vision (WSCG 2023).

This paper presents a framework and mobile video editing app for interactive artistic augmentation of human motion in videos. While creating motion effects with industry-standard software is time-intensive and requires expertise, and popular video effect apps have limited customization options, our approach enables a multitude of art-directable, hi...

Using software metrics as a method of quantification of software, various approaches were proposed for locating defect-prone source code units within software projects. Most of these approaches rely on supervised learning algorithms, which require labeled data for adjusting their parameters during the learning phase. Usually, such labeled training...

Offers to buy and sell cryptocurrencies on exchanges are collected in an order book as pairs of amount and price provided with a timestamp. Contrary to tick data, which only reflects the last transaction price on an exchange, the order book reflects the market's actual price information and the available volume. Until now, no system has been presen...

Most cryptocurrencies are bought and sold on centralized exchanges that manage supply and demand via an order book. Besides trading fees, the high liquidity of a market is the most relevant reason for choosing one exchange over the other. However, as the different liquidity measures rely on the order book, external events that cause people to sell...

Trading for a currency pair on centralized crypto exchanges is organized via an order book, which collects all open buy and sell orders at any given time and thus forms the basis for price formation. Usually, the exchanges provide basic visualizations, which show the accumulated buy and sell volume in an animated 2D representation. However, this vi...

Real estate is the largest asset class and is equally popular with professional and retail investors. However, this asset class has the disadvantage that it is very illiquid, and investments have a high entry barrier in terms of equity. The adoption of the Electronic Securities Act in 2021 by the German Bundestag has created the legal framework for...

Continuous Integration and Continuous Delivery are best practices used in the context of DevOps. By using automated pipelines for building and testing small software changes, possible risks are intended to be detected early. Those pipelines continuously generate log events that are collected in semi-structured log files. In practice, these log fil...

During the software development process, occurring problems are collected and managed as bug reports using bug tracking systems. Usually, a bug report is specified by a title, a more detailed description, and additional categorical information, e.g., the affected component or the reporter. It is the task of the triage owner to assign open bug repor...

More and more people are using images and videos as a communication tool. Often, such visual media is edited or stylized using software applications to become more visually attractive. The data that is produced by the editing process contains useful information on how users interact with the software and data yielding respective results. In this co...

Presentation of research paper "A Service-based Preset Recommendation System for Image Stylization Applications" at the "7th International Conference on Human Computer Interaction Theory and Applications (HUCAPP 2023)"

Self-Supervised Network Projections (SSNP) are dimensionality reduction algorithms that produce low-dimensional layouts from high-dimensional data. By combining an autoencoder architecture with neighborhood information from a clustering algorithm, SSNP intend to learn an embedding that generates visually separated clusters. In this work, we extend...

For various program comprehension tasks, software visualization techniques can be beneficial by displaying aspects related to the behavior, structure, or evolution of software. In many cases, the question is related to the semantics of the source code files, e.g., the localization of files that implement specific features or the detection of files...