Deep Clustering for Unsupervised Learning of Visual Features
Citations
4,128 citations
2,942 citations
Cites methods from "Deep Clustering for Unsupervised Le..."
...DeepCluster [17] partially answers this question....
[...]
...While previous methods based on bootstrapping have used pseudo-labels [16], cluster indices [17] or a handful of labels [18, 19, 20], we propose to directly bootstrap the representations....
[...]
1,771 citations
Cites background or methods from "Deep Clustering for Unsupervised Le..."
...DeepCluster-v2 and SeLa-v2, which are obtained by applying various training improvements introduced in other self-supervised learning papers to DeepCluster [6] and SeLa [2]....
[...]
...Our work is also related to clustering-based methods [2, 3, 6, 7, 17, 25, 50, 53, 54, 59]....
[...]
...We also show that our multi-crop strategy is general, and improves the performance of different self-supervised methods, namely SimCLR [9], DeepCluster [6], and SeLa [2], between 2% and 4% top-1 accuracy on ImageNet....
[...]
...Typical clustering-based methods [2, 6] are offline in the sense that they alternate between a cluster assignment step where image features of the entire dataset are clustered, and a training step where the cluster assignments, i....
[...]
...For example, clustering-based methods discriminate between groups of images with similar features instead of individual images [6]....
[...]
1,733 citations
Cites background or methods from "Deep Clustering for Unsupervised Le..."
...Like contrastive learning, clustering-based methods require either a memory bank [5, 6, 1], large batches [7], or a queue [7] to provide enough samples for clustering....
[...]
...Clustering [5] is another way of avoiding constant output, and SwAV [7] incorporates online clustering into Siamese networks....
[...]
...Another category of methods for unsupervised representation learning are based on clustering [5, 6, 1, 7]....
[...]
1,453 citations
Cites methods from "Deep Clustering for Unsupervised Le..."
...As many previous unsupervised methods are evaluated with AlexNet [43] on ImageNet [16, 42, 17, 84, 54, 18, 85, 55, 22, 11, 83], we also include the the results of CMC using this network....
[...]
References
73,978 citations
55,235 citations
"Deep Clustering for Unsupervised Le..." refers methods in this paper
...Table 4 compares a VGG-16 and an AlexNet trained with DeepCluster on ImageNet and tested on the Pascal VOC 2007 object detection task with fine-tuning....
[...]
...[46] 42.3 58.0 DeepCluster 61.0 72.0 Table 5: mAP on instance-level image retrieval on Oxford and Paris dataset with a VGG-16....
[...]
...In the supervised setting, deeper architectures like VGG or ResNet [8] have a much higher accuracy on ImageNet than AlexNet....
[...]
...Table 5 reports the performance of a VGG-16 trained with different approaches obtained with Sobel filtering, except for Doersch et al ....
[...]
...Replacing an AlexNet by a VGG [30] significantly improves the quality of the features and their subsequent transfer performance....
[...]
49,914 citations
49,639 citations
"Deep Clustering for Unsupervised Le..." refers background or methods in this paper
...We train DeepCluster on the training set of ImageNet [12] (1, 281, 167 images distributed uniformly into 1, 000 classes)....
[...]
...The existence of ImageNet [12], a large fully-supervised dataset, has been fueling advances in pre-training of convnets....
[...]
42,067 citations