HD-CNN: Hierarchical Deep Convolutional Neural Networks for Large Scale Visual Recognition

By Zhicheng Yan, Robinson Piramuthu, Vignesh Jagadeesh, Dennis DeCoste, Wei Di, Yizhou Yu. 2014

483 citations recorded by the source corpus.

Subfields: Artificial Intelligence, 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 Artificial Intelligence

1. Mohit Bansal
2. Bernt Schiele
3. Trevor Darrell
4. Quoc V. Le
5. Yixuan Li
6. Jiashi Feng
7. Jinwoo Shin
8. Rogério Schmidt Feris

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Dacheng Tao
3. Ming-Hsuan Yang
4. Xiaogang Wang
5. Trevor Darrell
6. Bernt Schiele
7. Jitendra Malik
8. Andrew Zisserman

Citations in the corpus

• ALICE: Active Learning with Contrastive Natural Language Explanations (2020, methodology): “Yan et al. (2015) decompose classification task into a coarse category classification and a fine category classifica- tion.”
• Dynamic Neural Networks: A Survey (2021, background): “For instance, a small CNN is first executed to classify each sample into coarse categories, and specific sub-networks are conditionally activated based on the coarse predictions [103].”
• Tree-CNN: A hierarchical Deep Convolutional Neural Network for incremental learning (2018, background): “Ima-geNet [6] contains 1.2 million images with 1000 categories), supervised learning has become the leading paradigm in training DCNNs for image recognition.”
• Fully-Adaptive Feature Sharing in Multi-Task Networks with Applications in Person Attribute Classification (2016, methodology): “We note that other techniques such as HD-CNN [46] and Network of Experts [1] also group related classes to perform hierarchical classification, but these methods are not applicable for the multi-label setting (where labels are not mutually exclusive).”
• The iNaturalist Species Classification and Detection Dataset (2017, background): “This direction would encourage new error measures that incorporate taxonomic rank [25, 47].”
• Do Convolutional Neural Networks Learn Class Hierarchy? (2017, result): “Our results show more significant improvement on the ILSVRC dataset than HD-CNN, a recently proposed approach to implement hierarchical CNNs [72].”
• Tree-CNN: A Deep Convolutional Neural Network for Lifelong Learning (2018, background): “D EEP Convolutional Neural Networks (DCNNs) haveemerged as the leading architecture for large scale image classification in recent years [1].”
• Unsupervised Joint Mining of Deep Features and Image Labels for Large-Scale Radiology Image Categorization and Scene Recognition (2017, background): “The modeled category hierarchy can alleviate the highly uneven visual separability among discovered image categories [63].”
• Decoding the Deep: Exploring Class Hierarchies of Deep Representations Using Multiresolution Matrix Factorization (2017, background): “The networks are not trained to learn the hierarchy of categories (unlike the alternate works from [33, 7, 2]) – the”
• Learning Layer-Skippable Inference Network (2020, methodology): “HD-CNN [16] learned a hierarchical structure of classes by grouping fine categories into coarse classes and embedded deep CNNs into the category hierarchy for better fine-grained prediction.”
• Interpretable Image Recognition with Hierarchical Prototypes (2019, methodology): “…and Torralba 2012), CNNs, (Ahmed and Torresani 2017; Redmon and Farhadi 2017; Zhu and Bain 2017; Zhao, Li, and Xing 2011; Kuang et al. 2018; Yan et al. 2015; Tamaazousti, Borgne, and Hudelot 2017; Roy, Panda, and Roy 2018), and recently the use of a CNN and RNN together (Guo et al. 2017).”
• Hierarchical Novelty Detection for Visual Object Recognition (2018, background): “Incorporating the hierarchical taxonomy for object classification has been investigated in the literature, either to improve classification performance [6, 34], or to extend the classification tasks to obtain more informative results [8, 36].”
• Learning More Universal Representations for Transfer-Learning (2017, methodology): “The methods of the second approach [1], [22], [32], [35], [55], [56] give an answer to the problem that in joint training, different kinds of features are undesirably mixed.”
• Deep Decision Network for Multi-class Image Classification (2016, methodology): “Fourthly, DDN uses multiple decision layers capable of making early decisions as opposed to using one probabilistic averaging layer and finally our method results on CIFAR-100 is better even without any data-augmentation and 10 average testing [12] tricks as used in [32].”