CityPersons: A Diverse Dataset for Pedestrian Detection

By Shanshan Zhang, Rodrigo Benenson, Bernt Schiele. 2017

920 citations recorded by the source corpus.

Convnets have enabled significant progress in pedestrian detection recently, but there are still open questions regarding suitable architectures and training data. We revisit CNN design and point out key adaptations, enabling plain FasterRCNN to obtain state-of-the-art results on the Caltech dataset. To achieve further improvement from more and better data, we introduce CityPersons, a new set of person annotations on top of the Cityscapes dataset. The diversity of CityPersons allows us for the first time to train one single CNN model that generalizes well over multiple benchmarks. Moreover, with additional training with CityPersons, we obtain top results using FasterRCNN on Caltech, improving especially for more difficult cases (heavy occlusion and small scale) and providing higher localization quality.

Subfields: Computer Vision and Pattern Recognition, Electrical and Electronic Engineering, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Electrical and Electronic Engineering

In Evaluation & Benchmarks

Citations in the corpus

Cascade R-CNN: High Quality Object Detection and Instance Segmentation (2019, methodology): “Finally,we further present the results of a larger set of experiments, performed on various popular generic/specific object detection datasets, including PASCAL VOC [12], KITTI [15], CityPerson [63] and WiderFace [60].”
nuScenes: A Multimodal Dataset for Autonomous Driving (2019, background): “Other datasets focus exclusively on pedestrian annotations on images [20, 25, 79, 24, 88, 23, 58].”
BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learning (2018, background): “Some of the datasets focus on particular objects such as pedes-trians [9, 39].”
Fusion of Multispectral Data Through Illumination-aware Deep Neural Networks for Pedestrian Detection (2018, background): “Pedestrian detection is a popular research topic within the field of computer vision in the past decades [29, 5, 8, 11, 10, 4, 41].”
Rethinking the Competition Between Detection and ReID in Multiobject Tracking (2020, methodology): “For a fair comparison, we use the same training datasets as the baseline tracker JDE [8], consisting of ETH [30], CityPerson [31], CalTech [32], MOT17 [33], CUDK-SYSU [34] and PRW [35].”
Joint Object Detection and Multi-Object Tracking with Graph Neural Networks (2021, methodology): “Following TR-MOT [27], we use a mix of six pedestrian detection datasets for pre-training, including CalTech [70], MOT17 [65], CUHK-SYSU [32], ETH [71], PRW [72], and CityPersons [73].”
Virtual to Real Adaptation of Pedestrian Detectors (2020, methodology): “CityPersons [20] is a real-world dataset for pedestrian detection more focused on self-driving applications.”
Object Detection in 20 Years: A Survey (2019, methodology): “CityPersons [62] 2017 Built based on CityScapes dataset [63].”
Learning Efficient Single-Stage Pedestrian Detectors by Asymptotic Localization Fitting (2018, methodology): “Comparison with the state-of-the-art on the CityPersons[44].”
PIE: A Large-Scale Dataset and Models for Pedestrian Intention Estimation and Trajectory Prediction (2019, background): “Publicly available pedestrian detection datasets [9, 12, 44] can potentially be used for such a purpose, however, they lack necessary characteristics such as ego-vehicle information [9], temporal correspondence [44], or enough pedestrians samples with long tracks [12].”
Occlusion-aware R-CNN: Detecting Pedestrians in a Crowd (2018, methodology): “The CityPersons dataset [67] is built upon the semantic segmentation dataset Cityscapes [7] to provide a new dataset of interest for pedestrian detection.”
Adaptive NMS: Refining Pedestrian Detection in a Crowd (2019, methodology): “For two-stage detectors, we generally follow the adapted Faster R-CNN framework [47] and use the pre-trained VGG-16 [37] as the backbone.”
Detection in Crowded Scenes: One Proposal, Multiple Predictions (2020, methodology): “CityPersons [42] is one of the widely used benchmarks for pedestrian detection.”
A Simple Baseline for Pose Tracking in Videos of Crowed Scenes (2020, background): “9k image with 19k boxes) [27], CrowndHuman (15k images with 339k boxes) [18] and self-collected data (2k images with 30k boxes).”
Bi-box Regression for Pedestrian Detection and Occlusion Estimation (2018, methodology): “The CityPersons dataset [41] is a recently released pedestrian detection dataset which is diverse in terms of the numbers of countries, cities and seasons it covers.”
A Survey of Deep Active Learning (2020, methodology): “[4, 230] PRID [93], MARS [256], BDD100K [246], DukeMTMC-VideoReID [237], CityPersons [249], Caltech Pedestrian[52] Person Re-id”
BDD100K: A Diverse Driving Video Database with Scalable Annotation Tooling (2018, background): “Some of the datasets focus on particular objects such as pedestrians [9,25].”
Scale Match for Tiny Person Detection (2019, positive): “The scenarios of existing person/pedestrian benchmarks [2][6][24][5][4][8] e.g., CityPersons [27], are mainly in a near or middle distance.”