ImageNet: A large-scale hierarchical image database

By Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, K. Li, Li Fei-Fei. 2009

74,730 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Evaluation & Benchmarks

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 Evaluation & Benchmarks

Citations in the corpus

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding (2019, background): “Computer vision research has also demonstrated the importance of transfer learning from large pre-trained models, where an effective recipe is to fine-tune models pre-trained with ImageNet (Deng et al., 2009; Yosinski et al., 2014).”
Exploring the Limits of Language Modeling (2016, methodology): “We hope for future research to focus on reasonably sized datasets taking inspiration from recent advances seen in the computer vision community thanks to efforts such as Imagenet (Deng et al., 2009).”
Learned in Translation: Contextualized Word Vectors (2017, methodology): “In computer vision, convolutional neural networks (CNNs) pretrained on ImageNet [Krizhevsky et al., 2012, Deng et al., 2009] have become the de facto initialization for more complex and deeper models.”
Supervised Learning of Universal Sentence Representations from Natural Language Inference Data (2017, background): “Striking examples include face recognition (Taigman et al., 2014) and visual question answering (Antol et al., 2015), where image features trained on ImageNet (Deng et al., 2009) and word embeddings trained on large unsupervised corpora are combined.”
SQuAD: 100,000+ Questions for Machine Comprehension of Text (2016, background): “Historically, large, realistic datasets have played a critical role for driving ﬁelds forward—famous examples include ImageNet for object recognition (Deng et al., 2009) and the Penn Treebank for syntactic parsing (Marcus et al., 1993).”
How transferable are features in deep neural networks? (2014, methodology): “The ImageNet dataset, as released in the Large Scale Visual Recognition Challenge 2012 (ILSVRC2012) (Deng et al., 2009) contains 1,281,167 labeled training images and 50,000 test images, with each image labeled with one of 1000 classes.”
Object Detection via a Multi-region and Semantic Segmentation-Aware CNN Model (2015, positive): “For all the CNN models involved in our proposed system, we used the publicly available 16-layers VGG model [25] pre-trained on ImageNet [4] for the task of image classiﬁcation.”
Fast R-CNN (2015, methodology): “Following traditional R-CNN, we start from a ConvNet initialized with discriminative pre-training for ImageNet [4] classiﬁcation.”
Object Detection Networks on Convolutional Feature Maps (2015, methodology): “R-CNN [9], one particularly successful approach, starts with a pre-trained ImageNet [5] classiﬁcation network and then ﬁne-tunes the ConvNet, end-to-end, for detection.”
Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification (2015, background): “techniques [13, 30, 10, 36], aggressive data augmentation [18, 14, 29, 33], and large-scale data [4, 26].”
Convolutional neural networks at constrained time cost (2014, methodology): “Through the investigations, we obtain a model that achieves 11.8% top-5 error (10-view test) on ImageNet [3] and only takes 3 to 4 days training on a single GPU.”
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition (2014, methodology): “Using SPP-net, we compute the feature maps from the entire image only once, and then pool features in arbitrary regions (sub-images) to generate fixed-length representations for training the detectors.”
OverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks (2013, background): “Pre-training with extra data from the ImageNet Fall11 dataset improved this number to 11.2%.”
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation (2013, background): “The only class-specific computations are dot products between features and SVM weights and non-maximum suppression.”
Improving neural networks by preventing co-adaptation of feature detectors (2012, background): “ImageNet is an extremely challenging object recognition dataset consisting of thousands of highresolution images of thousands of classes of object [7].”
Intriguing properties of neural networks (2013, neutral): “• The ImageNet dataset [3].”