• Publications
  • Influence
The Cityscapes Dataset for Semantic Urban Scene Understanding
TLDR
We introduce Cityscapes, a benchmark suite and large-scale dataset to train and test approaches for pixel-level and instance-level semantic labeling. Expand
  • 3,863
  • 1010
  • PDF
CityPersons: A Diverse Dataset for Pedestrian Detection
TLDR
We introduce CityPersons, a new set of high quality bounding box annotations for pedestrian detection on the Cityscapes dataset (train, validation, and test sets). Expand
  • 294
  • 101
  • PDF
Learning Video Object Segmentation from Static Images
TLDR
We demonstrate that highly accurate object segmentation in videos can be enabled by using a convolutional neural network trained with static images only. Expand
  • 304
  • 88
  • PDF
Face Detection without Bells and Whistles
Face detection is a mature problem in computer vision. While diverse high performing face detectors have been proposed in the past, we present two surprising new top performance results. First, weExpand
  • 507
  • 86
  • PDF
What Makes for Effective Detection Proposals?
TLDR
We introduce a novel metric, the average recall (AR), which rewards both high recall and good localisation and correlates surprisingly well with detection performance. Expand
  • 559
  • 56
  • PDF
Ten Years of Pedestrian Detection, What Have We Learned?
TLDR
We analyse the remarkable progress of the last decade by dis- cussing the main ideas explored in the 40+ detectors currently present in the Caltech pedestrian detection benchmark. Expand
  • 601
  • 49
  • PDF
Simple Does It: Weakly Supervised Instance and Semantic Segmentation
TLDR
We show that when carefully designing the input labels from given bounding boxes, even a single round of training is enough to improve over previously reported weakly supervised results. Expand
  • 283
  • 49
  • PDF
Filtered channel features for pedestrian detection
TLDR
This paper starts from the observation that multiple top performing pedestrian detectors can be modelled by using an intermediate layer filtering low-level features in combination with a boosted decision forest. Expand
  • 317
  • 48
  • PDF
Pedestrian detection at 100 frames per second
TLDR
We present a new pedestrian detector that improves both in speed and quality over state-of-the-art. Expand
  • 580
  • 46
  • PDF
Traffic sign recognition — How far are we from the solution?
TLDR
We show that, without any application specific modification, existing methods for pedestrian detection, and for digit and face classification; can reach performances in the range of 95% ~ 99% of the perfect solution. Expand
  • 241
  • 32
  • PDF