Search Results for author: Oscar Beijbom

Found 16 papers, 10 papers with code

PolarStream: Streaming Object Detection and Segmentation with Polar Pillars

no code implementations NeurIPS 2021 Qi Chen, Sourabh Vora, Oscar Beijbom

Recent works recognized lidars as an inherently streaming data source and showed that the end-to-end latency of lidar perception models can be reduced significantly by operating on wedge-shaped point cloud sectors rather then the full point cloud.

Object object-detection +1

Multimodal Trajectory Prediction Conditioned on Lane-Graph Traversals

1 code implementation28 Jun 2021 Nachiket Deo, Eric M. Wolff, Oscar Beijbom

Accurately predicting the future motion of surrounding vehicles requires reasoning about the inherent uncertainty in driving behavior.

Decoder motion prediction +1

PolarStream: Streaming Lidar Object Detection and Segmentation with Polar Pillars

no code implementations14 Jun 2021 Qi Chen, Sourabh Vora, Oscar Beijbom

Recent works recognized lidars as an inherently streaming data source and showed that the end-to-end latency of lidar perception models can be reduced significantly by operating on wedge-shaped point cloud sectors rather then the full point cloud.

LIDAR Semantic Segmentation object-detection +1

PointPainting: Sequential Fusion for 3D Object Detection

4 code implementations CVPR 2020 Sourabh Vora, Alex H. Lang, Bassam Helou, Oscar Beijbom

Surprisingly, lidar-only methods outperform fusion methods on the main benchmark datasets, suggesting a gap in the literature.

3D Object Detection Object +5

Compact Bilinear Pooling

7 code implementations CVPR 2016 Yang Gao, Oscar Beijbom, Ning Zhang, Trevor Darrell

Bilinear models has been shown to achieve impressive performance on a wide range of visual tasks, such as semantic segmentation, fine grained recognition and face recognition.

Face Recognition Few-Shot Learning +3

Random Sampling in an Age of Automation: Minimizing Expenditures through Balanced Collection and Annotation

no code implementations26 Oct 2014 Oscar Beijbom

Methods for automated collection and annotation are changing the cost-structures of sampling surveys for a wide range of applications.

Efficient Large-Scale Structured Learning

no code implementations CVPR 2013 Steve Branson, Oscar Beijbom, Serge Belongie

Our method is shown to be 10-50 times faster than SVMstruct for cost-sensitive multiclass classification while being about as fast as the fastest 1-vs-all methods for multiclass classification.

Binary Classification Classification +3

Cannot find the paper you are looking for? You can Submit a new open access paper.