Attend Refine Repeat: Active Box Proposal Generation via In-Out Localization

By Spyridon Gidaris, Nikos Komodakis. 2016

82 citations recorded by the source corpus.

The problem of computing category agnostic bounding box proposals is utilized as a core component in many computer vision tasks and thus has lately attracted a lot of attention. In this work we propose a new approach to tackle this problem that is based on an active strategy for generating box proposals that starts from a set of seed boxes, which are uniformly distributed on the image, and then progressively moves its attention on the promising image areas where it is more likely to discover well localized bounding box proposals. We call our approach AttractioNet and a core component of it is a CNN-based category agnostic object location refinement module that is capable of yielding accurate and robust bounding box predictions regardless of the object category. We extensively evaluate our AttractioNet approach on several image datasets (i.e. COCO, PASCAL, ImageNet detection and NYU-Depth V2 datasets) reporting on all of them state-of-the-art results that surpass the previous work in the field by a significant margin and also providing strong empirical evidence that our approach is capable to generalize to unseen categories. Furthermore, we evaluate our AttractioNet proposals in the context of the object detection task using a VGG16-Net based detector and the achieved detection performance on COCO manages to significantly surpass all other VGG16-Net based detectors while even being competitive with a heavily tuned ResNet-101 based detector. Code as well as box proposals computed for several datasets are available at:: this https URL

Subfields: Artificial Intelligence, 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 Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Feature Pyramid Networks for Object Detection (2016, neutral): “7 - - - - - - - AttractioNet‡ [10] VGG16 + Wide ResNet§ 2016 53.”
Cascade R-CNN: High Quality Object Detection and Instance Segmentation (2019, methodology): “AttractioNet [17] VGG16+Wide ResNet 35.”
Learning Open-World Object Proposals without Learning to Classify (2021, background): “Recently, learning-based proposals [1], [2], [13], [23], [24] have demonstrated better performance than classical approaches in both precision and recall, and are an important part of two-stage detectors.”
Cascade R-CNN: Delving Into High Quality Object Detection (2017, methodology): “Some works [10, 11, 18] have argued that a single regression step of � is insufficient for accurate localization.”
Hybrid Task Cascade for Instance Segmentation (2019, methodology): “AttractioNet [15] introduces an Attend & Refine module to update bounding box locations iteratively.”
Acquisition of Localization Confidence for Accurate Object Detection (2018, background): “In a series of object detectors [31, 7, 6, 21], refined boxes will be fed to the bounding box regressor again and go through the refinement for another time.”
Scale-Aware Trident Networks for Object Detection (2019, unknown): “AttractionNet [15] VGG16 + Wide ResNet 35.”
Region Proposal by Guided Anchoring (2019, unknown): “10 AttractioNet [11] VGG-16 53.”
Cascade RPN: Delving into High-Quality Region Proposal Network with Adaptive Convolution (2019, methodology): “A number of studies have attempted to improve RPN by iterative refinement [14, 41].”
Spatial Memory for Context Reasoning in Object Detection (2017, background): “Another commonly used trick especially beneﬁcial for reducing localization error is iterative bounding box reﬁne-ment [25, 26, 69, 95], which leverages local image context to predict a better bounding box iteratively.”
D2Det: Towards High Quality Object Detection and Instance Segmentation (2020, background): “Several recent works have extended this framework by, for example, integrating pyramid representations [31, 28, 44, 5], extending to multi-stage detection [16, 2, 6, 24] and integrating a mask branch [19, 23, 34].”
Zoom Out-and-In Network with Map Attention Decision for Region Proposal and Object Detection (2017, methodology): “Definition of seen and unseen categories are the same with AttractioNet (Gidaris and Komodakis, 2016).”
Imbalance Problems in Object Detection: A Review (2019, methodology): “The authors show that such a cascaded scheme works better compared to the previous work, such as MultiRegion CNN [129] and AttractioNet [130], that iteratively apply the same network to the bounding boxes.”
Learning Chained Deep Features and Classifiers for Cascade in Object Detection (2017, methodology): “6% when using the AttractionNet [11] for region proposal.”