Jonas Uhrig
YOU?
Author Swipe
Neural Rendering for Sensor Adaptation in 3D Object Detection Open
Autonomous vehicles often have varying camera sensor setups, which is inevitable due to restricted placement options for different vehicle types. Training a perception model on one particular setup and evaluating it on a new, different sen…
Sparsity Invariant CNNs Open
In this paper, we consider convolutional neural networks operating on sparse inputs with an application to depth upsampling from sparse laser scan data. First, we show that traditional convolutional networks perform poorly when applied to …
View article: DeMoN: Depth and Motion Network for Learning Monocular Stereo
DeMoN: Depth and Motion Network for Learning Monocular Stereo Open
In this paper we formulate structure from motion as a learning problem. We\ntrain a convolutional network end-to-end to compute depth and camera motion\nfrom successive, unconstrained image pairs. The architecture is composed of\nmultiple …
View article: Joint Graph Decomposition and Node Labeling: Problem, Algorithms, Applications
Joint Graph Decomposition and Node Labeling: Problem, Algorithms, Applications Open
We state a combinatorial optimization problem whose feasible solutions define both a decomposition and a node labeling of a given graph. This problem offers a common mathematical abstraction of seemingly unrelated computer vision tasks, in…
View article: Joint Graph Decomposition and Node Labeling: Problem, Algorithms,\n Applications
Joint Graph Decomposition and Node Labeling: Problem, Algorithms,\n Applications Open
We state a combinatorial optimization problem whose feasible solutions define\nboth a decomposition and a node labeling of a given graph. This problem offers\na common mathematical abstraction of seemingly unrelated computer vision tasks,\…
Pixel-level Encoding and Depth Layering for Instance-level Semantic Labeling Open
Recent approaches for instance-aware semantic labeling have augmented convolutional neural networks (CNNs) with complex multi-task architectures or computationally expensive graphical models. We present a method that leverages a fully conv…