Edge Boxes: Locating Object Proposals from Edges

By C. Lawrence Zitnick, Piotr Doll'ar. 2014

2,859 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Citations in the corpus

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (2015, methodology): “EdgeBoxes [6] currently provides the best”
Object Detection via a Multi-region and Semantic Segmentation-Aware CNN Model (2015, methodology): “heir systems. We follow the same practice and we report results on tables7and8. Speciﬁcally, we trained our models on VOC 2007 and 2012 train+val datasets using both selective search [34] and EdgeBox [39] proposals. During test time we only use EdgeBox proposals that are faster to be computed. From ”
Object Detection Networks on Convolutional Feature Maps (2015, methodology): “shows improved accuracy if the region data source is augmented. A simple approach is to use two complementary object proposal algorithms. We augment the Selective Search proposal regions by EdgeBoxes [35]. We use the top-ranked 2,000 proposals generated by the default setting of EdgeBoxes (for a thr”
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition (2014, background): “Index Terms—Convolutional Neural Networks, Spatial Pyramid Pooling, Image Classification, Object Detection F”
Relation Networks for Object Detection (2017, background): “From the convolutional features and a sparse set of region proposals [48, 52, 38], a RoI pooling layer [24, 18, 23] extracts fixed resolution regional features (e.”
Webly Supervised Learning of Convolutional Networks (2015, methodology): “We again use EdgeBox [60] to propose regions of interest on images.”
Soft-NMS — Improving Object Detection with One Line of Code (2017, background): “In edge detection, NMS performs edge thinning to remove spurious responses [25, 1, 31].”
You Only Look Once: Unified, Real-Time Object Detection (2015, background): “These classifiers or localizers are run either in sliding window fashion over the whole image or on some subset of regions in the image [34, 15, 38].”
R-FCN: Object Detection via Region-based Fully Convolutional Networks (2016, methodology): “3 Related Work R-CNN [8] has demonstrated the effectiveness of using region proposals [28, 29] with deep networks.”
Training Region-Based Object Detectors with Online Hard Example Mining (2016, methodology): “R-CNN, in contrast, uses region proposals [1, 2, 3, 4, 6, 10, 18, 32, 34], a method that was made popular by the selective search algorithm [32].”
Learning to Refine Object Segments (2016, neutral): “rening the top 500 proposals per image (at which Learning to Rene Object Segments 13 Box Proposals Segmentation Proposals AR 10 AR 100 AR 1 K AUC AR 10 AR 100 AR 1 K AUC S AUC M AUC L AUC EdgeBoxes [34] 7.4 17.8 33.8 13.9 | | | | | | | Geodesic [36] 4.0 18.0 35.9 12.6 2.3 12.3 25.3 1.3 8.6 20.5 8.5 ”
Inside-Outside Net: Detecting Objects in Context with Skip Pooling and Recurrent Neural Networks (2015, background): “This matches our intuition that context features look broadly across the image, while multi-scale features capture more fine-grained details.”
Learning to Segment Object Candidates (2015, methodology): “These approaches usually fall into three categories: (1) objectness scoring [1, 34], in which proposals are extracted by measuring the objectness score of bounding boxes, (2) seed segmentation [14, 16, 17], where models start with multiple seed regions and generate separate foreground-background seg”
What Makes for Effective Detection Proposals? (2015, methodology): “We survey these approaches in §2.1 and §2.2, followed by an overview of alternate approaches in §2.3 and baselines in §2.4.”
Learning Open-World Object Proposals without Learning to Classify (2021, background): “In early works, the emphasis was on category-independent object proposals [21], [3], [22], [7], where the goal is to identify instances of all objects in the image irrespective of their category.”
Focal Loss for Dense Object Detection (2017, background): “, Selective Search [12], EdgeBoxes [13], DeepMask [14], [15], RPN [3]) rapidly narrows down the number of can-”
Object-Proposal Evaluation Protocol is ‘Gameable’ (2015, methodology): “The algorithms that fall under this category are [1, 4, 5, 10, 45, 48].”
Deep Learning Face Attributes in the Wild (2014, methodology): “We also investigate how these methods perform with respect to overlap ratio (IoU ), following [35, 21].”
Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering (2017, methodology): “In this work, rather than using hand-crafted or differentiable region proposals [41, 50, 17], we leverage Faster R-CNN [32], establishing a closer link between vision and language tasks and recent progress in object detection.”