Collecting a Large-scale Dataset of Fine-grained Cars

By Jonathan Krause, Jia Deng, Michael Stark, Li Fei-Fei. 2013

246 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

Do Better ImageNet Models Transfer Better? (2018, background): “These datasets covered a wide range of image classification tasks, including superordinate-level object classification (CIFAR-10 [37], CIFAR-100 [37], PASCAL VOC 2007 [19], Caltech-101 [20]); fine-grained object classification (Food-101 [5], Birdsnap [4], Stanford Cars [36], FGVC Aircraft [48], Oxfo”
AutoAugment: Learning Augmentation Policies from Data (2018, neutral): “Even on datasets for which ﬁne-tuning weights pre-trained on ImageNet does not help signiﬁcantly [27], e.g. Stanford Cars [28] and FGVC Aircraft [29], training with the ImageNet policy reduces test set error by 1.16% and 1.76%, respectively.”
How Well Does Self-Supervised Pre-Training Perform with Streaming Data? (2021, positive): “Following (Chen et al., 2020a), we consider 12 diverse image classiﬁcation datasets including Food-101 (Bossard et al., 2014), CIFAR10 (Krizhevsky et al., 2009), CIFAR100 (Krizhevsky et al., 2009), Birdsnap (Berg et al., 2014), SUN397 (Xiao et al., 2010), Standard Cars (Krause et al., 2013), FGVC Ai”
Domain Adaptive Transfer Learning with Specialist Models (2018, neutral): “In effect, ImageNet captures less of the discriminative factors of variation that is captured in either FGVC Aircraft and Stanford Cars.”
How Well Do Self-Supervised Models Transfer? (2020, methodology): “For many-shot recognition, we adopt the benchmark suite proposed in the transfer learning study [28], which includes the target datasets FGVC Air-craft [38], Caltech-101 [15], Stanford Cars [29], CIFAR-10 [30], CIFAR-100 [30], DTD [9], Oxford 102 Flowers [41], Food-101 [3], Oxford-IIIT Pets [43], SU”
How Well Self-Supervised Pre-Training Performs with Streaming Data? (2021, methodology): “Following [7], we consider 12 diverse image classiﬁcation datasets including Food-101 [4], CI-FAR10 [26], CIFAR100 [26], Birdsnap [3], SUN397 [58], Standard Cars [25], FGVC Aircraft [33], VOC2007 [14], DTD [10], Oxford-IIIT Pets [43], Caltech-101 [15] and Oxford 102 Flowers [39].”
Why Do Better Loss Functions Lead to Less Transferable Features? (2020, methodology): “We evaluated the transferability of the fixed features of our ImageNet-pretrained models by training linear or k-nearest neighbors (kNN) classifiers to classify 8 different natural image datasets: Food-101 [9], CIFAR-10 and CIFAR-100 [36], Birdsnap [6], SUN397 [77], Stanford Cars [35], Oxford-IIIT P”
Re-labeling ImageNet: from Single to Multi-Labels, from Global to Localized Labels (2021, methodology): “We evaluate ReLabelpretrained ResNet-50 on five fine-grained classification tasks: Food-101 [3], Stanford Cars [26], DTD [6], FGVC Aircraft [33], and Oxford Pets [35].”
Beyond Synthetic Noise: Deep Learning on Controlled Noisy Labels (2019, methodology): “We select the above architectures to be representative of diverse capacities, the accuracy of which on the Interestingly, the drop on the red noise is considerably smaller and even approaches zero on the Stanford Cars dataset.”
The Computational Limits of Deep Learning (2020, unknown): “…datasets, including Caltech 101 [111], Caltech 256 [112], ImageNet [24], CIFAR-10/100 [113], MNIST [114], SVHN [115], STL-10 [116], Fashion-MNIST [117], CINIC-10 [118], Flowers-102, iNaturalist [119], EMNIST-Letters [120], Kuzushiji-MNIST [121], Stanford Cars [122], and MultiMNIST [123].”
Vision-Language Models for Vision Tasks: A Survey (2023, background): “For example, the 27 image classification datasets cover a wide range of visual recognition tasks from the fine-grained tasks like Oxford-IIIT PETS [33] for pet identification and Stanford Cars [32] for car recognition, to the general tasks like ImageNet [52].”
Building a bird recognition app and large scale dataset with citizen scientists: The fine print in fine-grained dataset collection (2015, background): “Various domains have already had datasets introduced including Birds (the CUB-200 [35] and recently announced Birdsnap [2] datasets), Flowers [25, 1], Dogs and Cats [15, 27, 21], Stoneflies [24], Butterflies [19] and Fish [4] along with man-made domains such as Airplanes [23], Cars [16], and Shoes[3”
Divide and Contrast: Self-supervised Learning from Uncurated Data (2021, methodology): “We evaluate models pre-trained on ImageNet, YFCC100M and JFT-300M with a linear classifier on 12 downstream classification tasks: Food-101 [6], CIFAR-10/100 [54], Birdsnap [4], SUN397 [98], Stanford Cars [53], FGVC Aircraft [61], PASCAL VOC 2007 [27], Describable Textures (DTD) [19], Oxford-IIIT Pet”
AutoFormer: Searching Transformers for Visual Recognition (2021, methodology): “We transfer Autoformer to a list of commonly used recognition datasets: 1) general classiﬁcation: CIFAR-10 and CIFAR-100 [27]; 2) ﬁne-grained clas-siﬁcation: Stanford Car [26], FLowers [37] and Oxford-III Pets [38].”