MSc defense by Larissa Gorzawski

On Wednesday 5th of October at 2pm Larissa will present her M.Sc. thesis “Deep transfer learning on street-level imagery for classification of seismic building types in Lima, Peru”.

From the abstract: “Comprehensive exposure models for seismic risk assessment require accurate building inventories in the endangered areas. The learning capabilities of Deep Learning (DL) can be combined with street-level imagery to categorize buildings in an automated way. Since the training of a DL model requires large amounts of data, in this thesis, a transfer learning approach will be employed to adapt an already trained model to a new study area and a different image dataset while minimizing the labeling requirements. The used model was trained by Aravena Pelizari, et al. (2021) with Google Street View (GSV) images in Santiago de Chile and will be adapted to the Peruvian capital Lima and to street-level imagery of the open-source platform Mapillary.

Three data-driven active learning (AL) strategies are designed and implemented with a pre-labeled pool of images: an initial cluster-based sampling with subsequent margin sampling and two additional augmentation strategies exploring the impact of either the most similar source domain images or secondly the addition of high-confidence semi-labeled target domain images.

The methods could achieve an improvement of the F1 accuracy score from 0.31 to 0.67 with a comparatively small amount of labeled images. Though the methods did all per-form similarly, the initial clustering and the semi-labeling of additional target domain images were the most promising approaches. The class accuracies indicate that the class differences between the domains could be learned at least partly quite successfully, while the performance of the data-driven AL methods was presumably limited by the noisiness of the dataset. The methods provided promising first results and could be further improved with diversity-based batch sampling as well as an extension of the semi-supervised learning approach.”

supervisors: Hannes Taubenböck and Christian Geiss

read more news:

Internship Report on Tuesday, May 21st at 12:00

Internship Report on Tuesday, May 21st at 12:00

NEW DATE: On Tuesday, May 21st, Elly Schmid will present her internship at 12:00 in seminar room 3, John-Skilton-Str. 4a. : From the abstract: The internship was carried out as part of the HEATS-(Urban heat) Project of the Georisks team at the Earth Observation...

Spatial Earth Observation R packages by our EAGLEs

Spatial Earth Observation R packages by our EAGLEs

Our EAGLE students that took and passed our Introduction to spatial programming course had to submit an R package that applies spatial methods for a variety of Earth Observation data. We are very proud to show the huge diversity of very interesting and useful R...

publication by our EAGLE Ben Lee

publication by our EAGLE Ben Lee

Our EAGLE Ben conducted an internship during his EAGLE studies which led to a new publication on "Predicting resilience of migratory birds to environmental change" jointly with the Alfred Wegener Institute Helmholtz Centre for Polar and Marine Research, Section Polar...

Internship Report on Tuesday, April 30 at 14:00

Internship Report on Tuesday, April 30 at 14:00

On Tuesday, April 30 Konstantin Müller will present his internship " GDELT News Analysis of the Noto Earthquake via ERNIE" at 14:00 in 01.B.03, John-Skilton-Str. 4a. : From the abstract: The analysis of socioeconomic data has gained increasing importance. The exchange...