Joel Schlosser

2 accepted papers

2019

Multi-class classification without multi-class labels

Yen-Chang Hsu, Zhaoyang Lv, Joel Schlosser, Phillip Odom, Zsolt Kira

ICLR 2019poster

This work presents a new strategy for multi-class classification that requires no class-specific labels, but instead leverages pairwise similarity between examples, which is a weaker form of annotation. The proposed method, meta classification learning, optimizes a binary classifier for pairwise sim…

Cited by 215 Source ↗PDF ↗code ↗Scholar ↗

2016

Fusing LIDAR and images for pedestrian detection using convolutional neural networks

Joel Schlosser, Christopher K. Chow, Zsolt Kira

ICRA 2016

In this paper, we explore various aspects of fusing LIDAR and color imagery for pedestrian detection in the context of convolutional neural networks (CNNs), which have recently become state-of-art for many vision problems. We incorporate LIDAR by up-sampling the point cloud to a dense depth map and

Cited by 127 Source ↗Scholar ↗