Weakly Supervised Deep Detection Networks

By Hakan Bilen, Andrea Vedaldi. 2015

843 citations recorded by the source corpus.

Weakly supervised learning of object detection is an important problem in image understanding that still does not have a satisfactory solution. In this paper, we address this problem by exploiting the power of deep convolutional neural networks pre-trained on large-scale image-level classification tasks. We propose a weakly supervised deep detection architecture that modifies one such network to operate at the level of image regions, performing simultaneously region selection and classification. Trained as an image classifier, the architecture implicitly learns object detectors that are better than alternative weakly supervised detection systems on the PASCAL VOC data. The model, which is a simple and elegant end-to-end architecture, outperforms standard data augmentation and fine-tuning techniques for the task of image-level classification as well.

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

Learning Open-World Object Proposals without Learning to Classify (2021, background): “Weakly-supervised approaches [26], [27], [28] aim to utilize abundant image-level labels and leverage class-agnostic box proposals to build detectors.”
Extreme Clicking for Efficient Object Annotation (2017, methodology): “5 2007 WSOL: Bilen and Vedaldi [5] – – 54 35 35 0 0”
Incremental Few-Shot Object Detection (2020, neutral): “There are other techniques that aim to minimise the amount of data labelling for object detection such as weakly supervised learning [4, 10] and zero-shot learning [58, 37, 2].”
Open-Vocabulary Object Detection Using Captions (2020, methodology): “We compare to WSDDN [4], as well as Cap2Det [43] which better utilizes captions.”
Weakly-Supervised Learning of Visual Relations (2017, background): “Most of the work on weakly-supervised learning for visual recognition has focused on learning objects [3, 11, 35].”
Few-Shot Object Detection via Feature Reweighting (2018, background): “The weakly-supervised setting [3, 7, 36] considers the problem of training object detectors with only image-level labels, but without bounding box annotations, which are more expensive to obtain.”
Large-Scale Weakly-Supervised Pre-Training for Video Action Recognition (2019, background): “Learning from weak supervision: Given the known challenges in gathering exhaustive annotations for various image and video tasks, leveraging object labels and other meta information to supply weak supervision [7, 15, 19, 22, 30, 31, 39, 40, 50, 52, 53, 54, 55, 56, 59] is a widely-adopted approach.”
DetCLIPv2: Scalable Open-Vocabulary Object Detection Pre-training via Word-Region Alignment (2023, methodology): “On the other hand, Weakly-supervised Object Detection (WSOD) methods [3,10,36] aim to establish localization-capable detectors by leveraging image-level labels, which also require a set of pre-defined categories.”
Weakly Supervised Action Localization by Sparse Temporal Pooling Network (2017, methodology): “In particular, the formulation of UntrimmedNet for action localization heavily relies on the idea proposed in [2].”
See Better Before Looking Closer: Weakly Supervised Data Augmentation Network for Fine-Grained Visual Classification (2019, methodology): “Instead of proposing region of interest bounding boxes [1, 5], we represent object’s parts or visual pattern by attention maps which are generated by the convolutions.”
A Survey of Deep Learning-Based Object Detection (2019, neutral): “Weakly supervised deep detection network in [222] is a representative framework for weakly supervised object detec-”
Omni-DETR: Omni-Supervised Object Detection with Transformers (2022, neutral): “image-level tags [4, 17, 23, 42, 44] or instance-level points [8, 19, 32], and usually formulate WSOD as a multiple instance learning (MIL) problem.”
Few-Shot Object Detection and Viewpoint Estimation for Objects in the Wild (2020, background): “Object detection with limited labeled samples is also addressed by approaches targeting weak supervision [49, 5, 6, 47] and zero-shot learning [3, 40, 66], but these settings are different from ours.”
iCAN: Instance-Centric Attention Network for Human-Object Interaction Detection (2018, background): “aims to recognize interactions (if any) between all pairs of person and object instances. Visual relationship detection. A number of recent work addresses the problem of detecting visual relationship [1,7,19,22,23,32,43,44] and generating scene graph [25,39,42]. Several papers leverage some forms of”
Deep Learning for Generic Object Detection: A Survey (2018, methodology): “rvised learning is not scalable in the absence of fully labelled training data, therefore it is valuable to study how the power of CNNs can be leveraged in weakly supervised or unsupervised detection [15,45,187]. (8) 3D Object Detection: The progress of depth cameras has enabled the acquisition of d”
Weakly Supervised Object Localization and Detection: A Survey (2021, positive): “While other approaches, e.g., [8], [110], [157], focus on developing weakly supervised learning frameworks for unconstrained object categories, i.e., learning frameworks can be extended to learn object detector for the given category-speciﬁc weakly-labelled training images.”
UntrimmedNets for Weakly Supervised Action Recognition and Detection (2017, methodology): “Weakly supervised learning was extensively studied in object recognition and detection [32, 1, 9, 4], and there were several works adapting this method to learn action models from videos [26, 8, 2, 3, 15, 24].”
Object Detection in 20 Years: A Survey (2019, neutral): “In addition to the above approaches, some other researchers considered the WSOD as a proposal ranking process by selecting the most informative regions and then training these regions with image-level annotation [213].”