Caltech-UCSD Birds 200

By Peter Welinder, Steve Branson, Takeshi Mita, Catherine Wah, Florian Schroff, Serge J. Belongie, Pietro Perona. 2010

1,411 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Evaluation & Benchmarks

Citations in the corpus

Microsoft COCO: Common Objects in Context (2014, methodology): “Unlike previous datasets containing entry-level categories [29], such as “dog” or “chair,” like [28], ImageNet used the WordNet Hierarchy [30] to obtain both entry-level and ﬁne-grained [31] categories.”
Exploring the Limits of Weakly Supervised Pretraining (2018, methodology): “between training and test sets. Overlap can exists because images may come from the same underlying sources (e.g., Wikipedia, Flickr, Google). For instance, ∼5% of the images in the val-CUB-6k200 set [21] also appear in train-IN-1M-1k, and 1.78% of images in val-IN-50k-1k set are in the JFT-300Mtrai”
Predicting Deep Zero-Shot Convolutional Neural Networks Using Textual Descriptions (2015, methodology): “ The training time for the fully connected model is 1-2 hours on a GTX Titan, whereas the joint fc+conv model takes 4 hours to train. 5.2. Caltech UCSD Bird The 200-category Caltech UCSD bird dataset [28] is one of the most widely used and competitive ﬁne-grained classiﬁcation benchmarks. We evaluat”
Generative Adversarial Text to Image Synthesis (2016, methodology): “These approaches exceed the previous stateof-the-art using attributes for zero-shot visual recognition on the Caltech-UCSD birds database (Wah et al., 2011), and also are capable of zero-shot caption-based retrieval.”
DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition (2013, methodology): “…visual recognition challenges have tasks with few training examples; e.g., when a user is deﬁning a category “on-the-ﬂy” us-ing speciﬁc examples, or for ﬁne-grained recognition challenges (Welinder et al., 2010), attributes (Bourdev et al., 2011), and/or domain adaptation (Saenko et al., 2010).”
Scaling Vision Transformers (2021, positive): “Similar trends on other datasets, Oxford IIIT Pets [27], CIFAR-100 [23], and Caltech-UCSD Birds [46] are presented in the Appendix, Figure 9.”
Prototypical Networks for Few-shot Learning (2017, methodology): “In order to assess the suitability of our approach for zero-shot learning, we also run experiments on the Caltech-UCSD Birds (CUB) 200-2011 dataset [34].”
Towards fairer datasets: filtering and balancing the distribution of the people subtree in the ImageNet hierarchy (2019, background): “, the PASCAL dataset has only 20 classes [26], CelebA has 40 attributes [50], COCO has 80 object categories [48], the fine-grained CUB-200 dataset has 200 bird species [77].”
Learning Deep Features for Discriminative Localization (2015, methodology): “In this section, we apply our generic localizable deep features to identifying 200 bird species in the CUB-2002011 [27] dataset.”
Part-Based R-CNNs for Fine-Grained Category Detection (2014, methodology): “We present state-of-the-art results evaluating our approach on the widely used ﬁne-grained benchmark Caltech-UCSD bird dataset [42].”
Learning Deep Representations of Fine-Grained Visual Descriptions (2016, methodology): “All neural text model architectures are the same as we used for CUB, and we used the same hyperparameters from crossvalidation on CUB.”
Inferring Semantic Layout for Hierarchical Text-to-Image Synthesis (2018, background): “, which is much more complicated than generating a single, large object as in simpler datasets [33, 17].”
Variational Information Distillation for Knowledge Transfer (2019, background): “, knowledge from natural image classification [23] is likely to be useful for fine-grained bird classification [29].”
The application of two-level attention models in deep convolutional neural network for fine-grained image classification (2014, methodology): “Fine-grained classification is to recognize subordinatelevel categories under some basic-level category, e.g., classifying different bird types [20], dog breeds [11], flower species [14], aircraft models [13] etc.”
The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition (2015, background): “Considerable work has also gone into the challenging task of curating finegrained datasets [4,58,27,30,31,59,65,60,70] and developing interactive methods for recognition with a human in the loop [6,62,61,63].”
This Looks Like That, Because ... Explaining Prototypes for Interpretable Image Recognition (2020, methodology): “For our experiments, we use the Caltech-UCSD Birds dataset [45], a data set for bird species identification which was also used by Chen et al.”
FearNet: Brain-Inspired Model for Incremental Learning (2017, background): “CUB-200 is a fine-grained image classification dataset containing high resolution images of 200 different bird species (Welinder et al., 2010).”
Anti-Retroactive Interference for Lifelong Learning (2022, methodology): “For CUB200 and ImageNet100, each incremental task is trained for 100 epochs.”
Probing Representation Forgetting in Supervised and Unsupervised Continual Learning (2022, methodology): “Finally though our work studies a diverse task sequence ImageNet → Scenes → CUB → Flowers, to fully understand the behavior of representation forgetting, results over more distant tasks may be needed (e.g. ImageNet → Sketch Images [13,44])”