Fast Feature Pyramids for Object Detection

By Piotr Doll'ar, Ron Appel, Serge J. Belongie, Pietro Perona. 2014

2,014 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

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 Media Technology

Citations in the corpus

Feature Pyramid Networks for Object Detection (2016, background): “ed in numerous works for image classiﬁcation, object detection, human pose estimation, and more. There has also been signiﬁcant interest in computing featurized image pyramids quickly. Doll´ar et al. [6] demonstrated fast pyramid computation by ﬁrst computing a sparsely sampled (in scale) pyramid an”
A MultiPath Network for Object Detection (2016, methodology): “Until recently, the sliding window paradigm was dominant [6, 32], especially for face and pedestrian detection.”
Synthetic Data for Text Localisation in Natural Images (2016, methodology): “The pipeline combines general purpose features such as HoG [6], EdgeBoxes [48] and Aggregate Channel Features [7] and brings in text speciﬁc (CNN) features only in the later stages, where patches are ﬁnally recognised as speciﬁc words.”
CoupleNet: Coupling Global Structure with Local Parts for Object Detection (2017, methodology): “ll three challenging datasets, i.e. a mAP of 82:7% on VOC07, 80:4% on VOC12, and 34:4% on MS COCO. 2. Related work Before the arrival of CNN, visual tasks have been dominated by traditional paradigms [4,6,23,26,24]. As one of an outstanding framework, DPM [6] described the object system using mixtur”
A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection (2016, methodology): “ctors, based on the sliding window paradigm, search for objects at multiple scales and aspect ratios. While real-time detectors are available for certain classes of objects, e.g. faces or pedestrians [1,2], it has proven diﬃcult to build detectors of multiple object classes under this paradigm. Rece”
Data Distillation: Towards Omni-Supervised Learning (2017, neutral): “, [45, 8, 7, 37]).”
Acquisition of Localization Confidence for Accurate Object Detection (2018, methodology): “The refinement further improves AP90 by 2.8% and the overall AP by 0.8% even for Cascade R-CNN which has a three-stage bounding box regressor.”
Reading Text in the Wild with Convolutional Neural Networks (2014, methodology): “We use a combination of EdgeBox proposals (Zitnick andDollár 2014) and a trained aggregate channel features detector (Dollár et al. 2014) to generate candidate word bounding boxes.”
HyperNet: Towards Accurate Region Proposal Generation and Joint Object Detection (2016, methodology): “ 100 proposals and (4) classiﬁes and makes adjustment for each region. Object detection aims to localize and recognize every object instance with a bounding box [9][30]. The DPM [10] and its variants [7][3] have been the dominating methods for years. These methods use image descriptors such as HOG [”
MOT16: A Benchmark for Multi-Object Tracking (2016, neutral): “valuation metrics that we provide in our benchmark. 7 Recall 0 0.5 1 Precision 0 0.2 0.4 0.6 0.8 1 Rcll: 24.2 % Prc: 65.7 % Rcll: 22.0 % Prc: 60.1 % Detector Performance Training Set Test Set (a) ACF [11] Recall 0 0.2 0.4 0.6 0.8 1 Precision 0 0.2 0.4 0.6 0.8 1 Rcll: 41.2 % Prc: 17.5 % Rcll: 47.2 % ”
Deep Feature Pyramid Reconfiguration for Object Detection (2018, background): “Despite the ineﬃciency, this kind of approaches have been applied for object detection and many other tasks along with hand-engineered features [7,12].”
Detecting Masked Faces in the Wild with LLE-CNNs (2017, methodology): “[29] proposed a face detector that utilized the feature aggregation framework [3], while the features were generated via CNN.”
CrowdHuman: A Benchmark for Detecting Human in a Crowd (2018, methodology): “Detecting people in images is among the most important components of computer vision and has attracted increasing attention in recent years [29, 14, 32, 30, 10, 5, 4, 6, 18].”
Multispectral pedestrian detection: Benchmark dataset and baseline (2015, background): “Pedestrian detection is an active research area in the field of computer vision, since it is an essential and significant task for a surveillance or tracking system [6] [14] [24] [27] [28], as well as for pedestrian safety [9] [13] [15].”
Multi-view Face Detection Using Deep Convolutional Neural Networks (2015, methodology): “This is interesting as many of the other methods such as [22, 2] requires a significantly larger number of resizing per octave, e.g. 8.”
Learning Chained Deep Features and Classifiers for Cascade in Object Detection (2017, methodology): “Cascade can be applied for SVM [19, 8], boosted classi-ﬁers [6, 17, 31], and ConvNets [33].”
Learning Efficient Single-Stage Pedestrian Detectors by Asymptotic Localization Fitting (2018, methodology): “In TA-CNN [38], the ACF detector [10] is employed for proposal generation, then pedestrian detection is jointly optimized with an auxiliary semantic task.”
3D Object Proposals Using Stereo Imagery for Accurate Object Class Detection (2016, methodology): “The recently proposed 3DVP [41] employs ACF detectors [42] and learns occlusion patterns with 3D voxels.”