Publications de Radhakrishna Achanta

Exploiting the Signal-Leak Bias in Diffusion Models

Sabine Süsstrunk, Radhakrishna Achanta, Mahmut Sami Arpa, Martin Nicolas Everaert, Athanasios Fitsios

There is a bias in the inference pipeline of most diffusion models. This bias arises from a signal leak whose distribution deviates from the noise distribution, creating a discrepancy between training and inference processes. We demonstrate that this signa ...

2024

Partial annotations in active learning for semantic segmentation

Katrin Beyer, Radhakrishna Achanta

Semantic segmentation with deep learning plays a crucial role in various fields, including civil engineering, particularly in tasks such as damage assessment and urban planning. This paper addresses the challenge of efficiently training deep learning model ...

2024

Digital Twins of Stone Masonry Buildings for Damage Assessment

Katrin Beyer, Radhakrishna Achanta, Bryan German Pantoja Rosero

Digital twins are virtual models of physical objects or systems that enable real-time monitoring and analysis. In the ﬁeld of stone masonry buildings, digital twins can be used to assess damage, predict maintenance needs, and opti- mize building performanc ...

Springer2024

Diffusion in Style

Sabine Süsstrunk, Radhakrishna Achanta, Mahmut Sami Arpa, Martin Nicolas Everaert

We present Diffusion in Style, a simple method to adapt Stable Diffusion to any desired style, using only a small set of target images. It is based on the key observation that the style of the images generated by Stable Diffusion is tied to the initial lat ...

2023

Damage-augmented digital twins towards the automated inspection of buildings

Katrin Beyer, Radhakrishna Achanta, Bryan German Pantoja Rosero

Current procedures for the rapid inspection of buildings and infrastructure are subjective, time-consuming, and cumbersome to document, necessitating new technologies to automate the process and eliminate these shortcomings. Fortunately, recent development ...

2023

Towards real-time digital twins for post-earthquake damage assessment of masonry buildings

Katrin Beyer, Radhakrishna Achanta, Bryan German Pantoja Rosero

Currently, post-earthquake assessment is performed visually, which implies that inspectors visit the damaged buildings and, based on their expertise, assign a damage grade. Yet, visual inspection is subjective, time-consuming and difficult to document. In ...

2023

Automated image-based generation of finite element models for masonry buildings

Katrin Beyer, Radhakrishna Achanta, Bryan German Pantoja Rosero

To predict the response of masonry buildings to various types of loads, engineers use finite element models, specifically solid-element and macro-element models. For predicting masonry responses to seismic events in particular, equivalent frame models-a su ...

SPRINGER2023

VETIM: Expanding the Vocabulary of Text-to-Image Models only with Text

Sabine Süsstrunk, Radhakrishna Achanta, Mahmut Sami Arpa, Martin Nicolas Everaert

Text-to-image models, such as Stable Diffusion, can generate high-quality images from simple textual prompts. With methods such as Textual Inversion, it is possible to expand the vocabulary of these models with additional concepts, by learning the vocabula ...

BMVA2023

Digital twins for post-earthquake damage assessment of free-standing buildings including damage information

Katrin Beyer, Radhakrishna Achanta, Bryan German Pantoja Rosero, Amir Rezaie

We present a pipeline for the automatic generation of digital twins of free-standing buildings for damage assessment. Our method takes multiple view images of the target building as input and processes them to produce a 3D model in levels of detail (LOD) f ...

2022

Generating LOD3 building models from structure-from-motion and semantic segmentation

Pascal Fua, Katrin Beyer, Radhakrishna Achanta, Bryan German Pantoja Rosero, Mateusz Kozinski, Fernando Perez Cruz

This paper describes a pipeline for automatically generating level of detail (LOD) models (digital twins), specifically LOD2 and LOD3, from free-standing buildings. Our approach combines structure from motion (SfM) with deep-learning-based segmentation tec ...

2022