Is object localization for free? - Weakly-supervised learning with convolutional neural networks

By Maxime Oquab, L. Eon Bottou, Ivan Laptev, Josef Sivic. 2015

934 citations recorded by the source corpus.

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

The Cityscapes Dataset for Semantic Urban Scene Understanding (2016, background): “Overall, we believe that the unique characteristics of our dataset (e.g. scale variation, amount of small objects, focus on urban street scenes) allow for more such novel insights.”
Learning to Segment Object Candidates (2015, background): “Moreover, these models learn sufficiently general image features, which can be transferred to many different tasks [10, 11, 3, 22, 23].”
Deep Learning Face Attributes in the Wild (2014, background): “Deep learning [18, 34, 23, 7, 19, 32, 31, 13, 33, 22, 3, 28] recently achieved great success in attribute prediction, due to their ability to learn compact and discriminative features.”
Learning Sight from Sound: Ambient Sound Provides Supervision for Visual Learning (2017, positive): “To help understand whether the convolutional units considered in Section 4 directly convey semantics, we also created a global max-pooling feature (similar to Oquab et al (2015)), where we applied max pooling over the entire convolutional layer.”
Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization (2016, methodology): “Others have investigated similar methods using global max pooling [34] and log-sum-exp pooling [35].”
UberNet: Training a Universal Convolutional Neural Network for Low-, Mid-, and High-Level Vision Using Diverse Datasets and Limited Memory (2016, background): “Certain recent works such as [20, 77, 78] manage to im-pute missing data in an EM-type approach, by exploiting domain-speciﬁc knowledge - e.g. by requesting that a ﬁxed percentage of the pixels contained in the bounding box of an object obtain the same label as the object.”
Objects that Sound (2017, methodology): “Our approach has similarities with [22] and [36] who used max and average pooling, respectively, to learn object detectors without bounding box annotations in the single visual modality setting, but use ImageNet pretrained networks and image-level labels.”
Learning Deep Features for Discriminative Localization (2015, methodology): “Recent work has shown that despite being trained on image-level labels, CNNs have the remarkable ability to localize objects [1, 16, 2, 15].”
Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer (2016, positive): “It was also exploited in computer-vision-related tasks such as image captioning Xu et al. (2015), visual question answering Yang et al. (2015), as well as in weakly-supervised object localization Oquab et al. (2015) and classiﬁcation Mnih et al. (2014), to mention a few characteristic examples.”
Towards Open Set Deep Networks (2015, background): “Thus, other uses of the OpenMax rejection can be to improve training process and aid in developing better localization techniques [30, 15].”
Visual Feature Attribution Using Wasserstein GANs (2017, methodology): “Pinpointing all evidence of a class is important for a variety of tasks such as weakly supervised localisation or segmentation of structures [41, 43, 65], and better understanding disease effects, and physiological or pathological processes in medical images [67, 18, 12, 19, 13, 27, 54, 30, 31, 63].”
Growing Interpretable Part Graphs on ConvNets via Multi-Shot Learning (2016, methodology): “Given image-level labels (without object bounding boxes), object discovery (Oquab et al. 2015; Simon and Rodner 2015; Pathak, Kr ahenb uhl, and Darrell 2015) can be achieved by identifying common foreground patterns from noisy background.”
Discovering Causal Signals in Images (2016, methodology): “The experiments described in this paper depend crucially on the properties of the features computed by the convolutional layers of a CNN [14].”
Not Using the Car to See the Sidewalk — Quantifying and Controlling the Effects of Context in Classification and Segmentation (2018, methodology): “Our classifier achieves similar mAP on real COCO data as reported in [14], with our mAP slightly lower (0.”
Top-Down Neural Attention by Excitation Backprop (2016, methodology): “Therefore, the comparison with CAM encompasses the comparison with the dense evaluation approach for weakly supervised localization [21].”
Ambient Sound Provides Supervision for Visual Learning (2016, methodology): “1 directly convey semantics, we also created a global max-pooling feature (similar to [29]), where we applied max pooling over the entire convolutional layer.”
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].”
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.”