Mid-level Visual Element Discovery as Discriminative Mode Seeking

By Carl Doersch, Abhinav Kumar Gupta, Alexei A. Efros. 2013

273 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

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 Representation & Self-Supervised Learning

Citations in the corpus

Unsupervised Visual Representation Learning by Context Prediction (2015, background): “Finally, our work is related to a line of research on discriminative patch mining [13, 47, 26, 34, 49, 11], which has emphasized weak supervision as a means of object discovery.”
Learning to See by Moving (2015, background): “This hypothesis inspired research in fully unsupervised learning from visual imagery and led to popular algorithms such as ICA [5, 20], autoencoders [24, 9], sparse coding [32], restricted boltzmann machines [27, 37] and others [6].”
Unsupervised Learning of Visual Representations Using Videos (2015, neutral): “org/patients-and-public/good-vision-throughoutlife/childrens-vision/infant-vision-birth-to-24-months-of-age selves [41, 6, 7].”
Deep Image: Scaling up Image Recognition (2015, methodology): “ow more examples to the model. We achieve accuracy of 82.4% using multi-scale testing. Table 6. Accuracy comparison of different methods on MIT-67 indoor scenes Method Accuracy(%) IFV [26] 60.8 MLrep [15] 64.0 CNN-SVM [40] 58.4 CNNaug-SVM [40] 69.0 CNN(AlexConvNet)+multiscale pooling [40] 68.9 MOP-C”
Transitive Invariance for Self-Supervised Visual Representation Learning (2017, background): “Our work is also closely related to mid-level patch clustering [46, 7, 8] and unsupervised discovery of semantic classes [43, 47] as we attempt to find reliable clusters in our affinity graph.”
Context as Supervisory Signal: Discovering Objects with Predictable Context (2014, methodology): “Standard clustering algorithms like K-means rely on a good distance metric, but unfortunately, distances in different regions of the feature space often aren’t comparable [14].”
Multi-scale Orderless Pooling of Deep Convolutional Activation Features (2014, background): “In fact, several recent methods achieving state-of-the-art results on this dataset are based on the idea of finding such patches [38,37,36].”
Learning Visual Question Answering by Bootstrapping Hard Attention (2018, background): “One line of work on discriminative patches builds a representation by selecting some patches and ignoring others, which has proved useful for object detection and classiﬁ-cation [56,57,58], and especially visualization [59].”
Single Image 3D without a Single 3D Image (2015, methodology): “Our approach uses visual elements discovered from a large dataset and draws from a rich literature on discriminative patch-discovery [37, 11, 10, 24, 3].”
Multi-scale Recognition with DAG-CNNs (2015, unknown): “Deep19-DAG 77.5 Deep19 [29] 70.8 Caffe-DAG 64.6 Caffe [14] 59.5 MOP-CNN [12] 68.9 Places [39] 68.2 Mid-level [6] 64.0 FV+BoP [15] 63.2 Disc.”
Deep multiple instance learning for image classification and auto-annotation (2015, methodology): “ROI+Gist [30] 26.1 DPM [28] 30.4 Object Bank [22] 37.6 RBow [29] 37.9 BoP [14] 46.1 miSVM [23] 46.4 D-Parts [37] 51.4 IFV [14] 60.8 MLrep [8] 64.0 CNN-SVM [34] needs a feature representation of length over 200, 000, and MLrep [8] employs the very timeconsuming fine tuning on the dataset.”
Incorporating Scene Context and Object Layout into Appearance Modeling (2014, background): “[17] and [2] learn scene parts in a jointly unsupervised/weakly-supervised manner.”