image/svg+xml Hello, My name is Tobias Pohlen. I am a machine learning and computer vision researcher.

Publications

Please see my CV for a detailed list of publications.

Full-Resolution Residual Networks for Semantic Segmentation in Street Scenes

Full-Resolution Residual Networks for Semantic Segmentation in Street Scenes

Semantic image segmentation is an essential component of modern autonomous driving systems, as an accurate understanding of the surrounding scene is crucial to navigation and action planning. Current state-of-the-art approaches in semantic image segmentation rely on pre-trained networks that were initially developed for classifying images as a whole. While these networks exhibit outstanding recognition performance (i.e., what is visible?), they lack localization accuracy (i.e., where precisely is something located?). Therefore, additional processing steps have to be performed in order to obtain pixel-accurate segmentation masks at the full image resolution. To alleviate this problem we propose a novel ResNet-like architecture that exhibits strong localization and recognition performance. We combine multi-scale context with pixel-level accuracy by using two processing streams within our network: One stream carries information at the full image resolution, enabling precise adherence to segment boundaries. The other stream undergoes a sequence of pooling operations to obtain robust features for recognition. The two streams are coupled at the full image resolution using residuals. Without additional processing steps and without pre-training, our approach achieves an intersection-over-union score of 71.8% on the Cityscapes dataset.

Learn more

Semantic Segmentation of Modular Furniture

Semantic Segmentation of Modular Furniture

This paper proposes an approach for the semantic segmentation and structural parsing of modular furniture items, such as cabinets, wardrobes, and bookshelves, into so called interaction elements. Such a segmentation into functional units is challenging not only due to the visual similarity of the different elements but also because of their often uniformly colored and low-texture appearance. Our method addresses these challenges by merging structural and appearance likelihoods of each element and jointly optimizing over shape, relative location, and class labels using Markov Chain Monte Carlo (MCMC) sampling. We propose a novel concept called rectangle coverings which provides a tight bound on the number of structural elements and hence narrows down the search space. We evaluate our approach's performance on a novel dataset of furniture items and demonstrate its applicability in practice.

Learn more

Projects

Sparse Johnson-Lindenstrauss Transform in MATLAB/MEX

Sparse Johnson-Lindenstrauss Transform in MATLAB/MEX

This is a C++/MEX implementation of the a sparse Johnson-Lindenstrauss transform (SJLT). The function allows you to efficiently sample an SJLT matrix in MATLAB.

Learn more

Kernel Ridge Regression Illustration

Seminar Paper on Random Sketches for KRR

Last semester I attended a seminar on compressive sensing. My topic was "Random Sketches for Kernel Ridge Regression".

Learn more

LibJunge Logo

LibJungle - A C++ Decision Jungle Library

As part of my recent seminar, I implemented a small C++ library in order to evaluate the performance of Decision Jungles. I release the code here under the BSD license.

Learn more

Decision Jungle Report Screenshot

Seminar Paper on Decision Jungles

In July 2014 I took a seminar on current topics in computer vision and machine learning with Prof. Bastian Leibe. My topic was Decision Jungles as introduced by J. Shotton et al. at NIPS 2013.

Learn more

Decision SVMs Screenshot

Seminar Paper on Support Vector Machines

In my fifth semester at RWTH Aachen University, I took a seminar on pattern and speech recognition with Prof. Herman Ney. I wrote a seminar paper and held a talk on support vector machines (SVMs).

Learn more

Fourier Descriptor Demo

Elliptic Fourier Descriptor

For a lab course I took in my fourth semester at RWTH Aachen University, we developed an optical character recognition system. As part of this lab course, I implemented a MATLAB demo to extract and visualize fourier shape descriptors.

Learn more

RWTH Latex Template

RWTH Latex Template

In my second semester at RWTH Aachen University I had to give my first seminar talk. It was about constraint logic programming in Prolog. Because I did not find any neat LaTeX template, I created my own.

Learn more