Fine-grained recognition without part annotations

By Jonathan Krause, Hailin Jin, Jianchao Yang, Li Fei-Fei. 2015

482 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Fine-Grained Image Classification via Combining Vision and Language (2017, methodology): “Further more, Krause et al. [18] only use object annotation at training stage to learn the part detectors, then localize the parts automatically in the testing stage.”
Recent advances in convolutional neural networks (2015, neutral): “In their latest paper [111].”
The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition (2015, background): “The majority of research in fine-grained recognition has focused on developing improved models for classification [1,3,5,7,9,8,14,16,18,20,21,22,28,29,36,37,41,42,49,51,50,66,68,69,71,73,72,76,77,75,78].”
This looks like that: deep learning for interpretable image recognition (2018, neutral): “; PA-CNN [19]: 82.8 (bb); MG-CNN [46]: 83.0 (bb), 81.7 (full); ST-CNN [16]: 84.1 (full); 2-level attn.”
Bilinear CNN Models for Fine-Grained Visual Recognition (2015, methodology): “A more up-to-date set of results can be found in [21] who recently reported excellent performance using on this dataset leveraging more accurate CNN models with a method to train part detectors in a weakly supervised manner.”
Compact Bilinear Pooling (2015, background): “This is, to the best of our knowledge, the state of the art performance on this dataset without part annotation [16, 23].”
The Devil is in the Tails: Fine-grained Classification in the Wild (2017, background): “Most ﬁne-grained research papers present a novel model for classiﬁcation [4, 6, 7, 12, 15, 17, 25, 31, 40, 45, 46, 57, 58, 59, 61, 62].”
Large Scale Fine-Grained Categorization and Domain-Specific Transfer Learning (2018, neutral): “[30] 82.”
Noisy Softmax: Improving the Generalization Ability of DCNN via Postponing the Early Softmax Saturation (2017, background): “Recently, deep convolutional neural networks (DCNNs) have taken the computer vision field by storm, significantly improving the state-of-the-art performances in many visual tasks, such as face recognition [43, 44, 33, 36], large-scale image classification [22, 39, 46, 11, 13], and fine-grained objec”
Multi-Attention Multi-Class Constraint for Fine-grained Image Recognition (2018, background): “…more recently proposed methods tend to localize object parts with weakly-supervised mechanisms, such as the combination of pose alignment and co-segmentation [8], dynamic spatial transformation of the input image for better alignment [21], and parallel CNNs for bilinear feature extraction [22].”
Object-Part Attention Model for Fine-Grained Image Classification (2017, methodology): “Most existing works rely on object or part annotations [6], [12], [13], while labeling is heavily labor consuming.”
SPDA-CNN: Unifying Semantic Part Detection and Abstraction for Fine-Grained Recognition (2016, methodology): “Other unsupervised methods [24, 25, 39, 34] could generate multiple object parts, but they are not guaranteed to produce small parts with semantic meanings.”
Learning to Navigate for Fine-grained Classification (2018, methodology): “Besides, Jonathan et al. [22] use co-segmentation and alignment to generate parts without part annotations but the bounding-box annotations are used during training.”
Generalized BackPropagation, Étude De Cas: Orthogonality (2016, methodology): “Our method, reaching the accuracy of 81.2%, is marginally worse compared to Krause et al. (2015) while again not beneﬁting from ﬁne-grained techniques.”
Mask-CNN: Localizing parts and selecting descriptors for fine-grained bird species categorization (2018, methodology): “1% Co-Segmentation [38] X VGG-19×2 126,976 82.”
Adapting Models to Signal Degradation using Distillation (2016, methodology): “3%) when a R-CNN based object detector is used to estimate bounding-boxes of objects at test time instead of using true bounding-boxes (Table 2 in [27], CNN+GT BBox+ft vs.”
Fine-Grained Image Analysis With Deep Learning: A Survey (2021, background): “, part constraints [72] or part alignment [66], to locate fine-grained parts.”