Learning and Transferring Mid-level Image Representations Using Convolutional Neural Networks

By Maxime Oquab, L. Eon Bottou, Ivan Laptev, Josef Sivic. 2014

3,302 citations recorded by the source corpus.

Subfields: Artificial Intelligence, 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 Artificial Intelligence

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

Citations in the corpus

Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition (2014, background): “9 method VOC 2007 Caltech101 VQ [15]† 56.07 74.41±1.0 LLC [18]† 57.66 76.95±0.4 FK [19]† 61.69 77.78±0.6 DeCAF [13] - 86.91±0.7 Zeiler & Fergus [4] 75.90‡ 86.5±0.5 Oquab et al. [34] 77.7 - Chatfield et al. [6] 82.42 88.54±0.3 ours 82.44 93.42±0.5 Table 8: Classification results for Pascal VOC 2007 (”
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer (2019, background): “computer vision [Oquab et al., 2014; Jia et al., 2014; Huh et al., 2016; Yosinski et al., 2014]) use a supervised dataset for pre-training, we were interested to see whether omitting the unsupervised task from the multi-task pre-training mixture still produced good results.”
Growing a Brain: Fine-Tuning by Increasing Model Capacity (2017, background): “Notable exceptions include [28], who introduce an adaptation layer to facilitate transfer.”
Fixing the train-test resolution discrepancy (2019, background): “Models pre-trained for image classification, usually on the ImageNet database [8], transfer to a variety of other applications [24].”
Revisiting Unreasonable Effectiveness of Data in Deep Learning Era (2017, background): “[28] showed that expanding the training data to cover 1512 labels from ImageNet-14M further improves the object detection performance.”
Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning (2020, background): “Learning good image representations is a key challenge in computer vision [1, 2, 3] as it allows for efficient training on downstream tasks [4, 5, 6, 7].”
Learning Visual Features from Large Weakly Supervised Data (2015, background): “We note that achieving state-of-the-art results [5, 34, 39, 60] on these datasets requires the development of tailored pipelines, e.g. , using many image transformations and model ensembles, which is out of the scope of this paper.”
Visualizing and Understanding Convolutional Networks (2013, methodology): “[21] who use a convnet with a more appropriate classifier.”
A critical analysis of self-supervision, or what we can learn from a single image (2019, background): “Self-supervision builds on the fact that convolutional neural networks (CNNs) transfer well between tasks (Shin et al., 2016; Oquab et al., 2014; Girshick, 2015; Huh et al., 2016).”
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks (2015, methodology): “Additionally, supervised CNNs trained on scene classiﬁcation learn object detectors (Oquab et al., 2014).”
Speeding up Convolutional Neural Networks with Low Rank Expansions (2014, background): “The increasing use of these networks for detection in sliding window approaches [9, 28, 33] and the desire to apply CNNs in real-world systems means the speed of inference becomes an important factor for applications.”
CrossTransformers: spatially-aware few-shot transfer (2020, background): “Deep neural network representations can bring some visual knowledge from datasets like ImageNet [67] to bear on different tasks beyond ImageNet [14, 31, 61], but empirically, this requires a non-trivial amount of labeled data in the new task.”
Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset (2017, methodology): “ be used for other tasks and in other domains. One of the early examples of this was using the fc7 features from a network trained on ImageNet for the PASCAL VOC classiﬁcation and detection challenge [10,23]. Furthermore, improvements in the deep architecture, changing from AlexNet to VGG-16, immedi”
CNN Features Off-the-Shelf: An Astounding Baseline for Recognition (2014, background): “And in response someone has quoted recent success stories [27, 16, 11] and someone else professed skepticism.”
Billion-scale semi-supervised learning for image classification (2019, background): “This approach is typically required for tasks for which no enough labels are available [29], such a ﬁne-grain classiﬁcation or detection (although a recent work [15] shows that learning from scratch is possible in the latter case).”
From captions to visual concepts and back (2014, background): “These features have been shown to generalize beyond just the ImageNet classification task and perform well for a variety of vision tasks [7, 12, 32].”