Harvesting Mid-level Visual Concepts from Large-Scale Internet Images

Li, Quannan; Wu, Jiajun; Tu, Zhuowen

doi:10.1109/cvpr.2013.115

Cited by 126 publications

(83 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Spatial Pooling Regions [52] 50.1% VC + VQ [53] 52.3% CNN-SVM [54] 58.4% Improved Fisher Vectors [55] 60.8% Mid Level Representation [56] 64.0% Multiscale Orderless Pooling [57] 68.9% TABLE VIII: Comparisons of our approach with the state-of-the-art class-imbalance approaches. The experimental protocols used for each dataset are shown in Fig.…”

Section: Results and Comparisonsmentioning

confidence: 99%

Cost-Sensitive Learning of Deep Feature Representations From Imbalanced Data

Khan

Hayat

Bennamoun

et al. 2018

IEEE Trans. Neural Netw. Learning Syst.

589

View full text Add to dashboard Cite

Abstract-Class imbalance is a common problem in the case of real-world object detection and classification tasks. Data of some classes is abundant making them an over-represented majority, and data of other classes is scarce, making them an underrepresented minority. This imbalance makes it challenging for a classifier to appropriately learn the discriminating boundaries of the majority and minority classes. In this work, we propose a cost-sensitive deep neural network which can automatically learn robust feature representations for both the majority and minority classes. During training, our learning procedure jointly optimizes the class-dependent costs and the neural network parameters. The proposed approach is applicable to both binary and multiclass problems without any modification. Moreover, as opposed to data level approaches, we do not alter the original data distribution which results in a lower computational cost during the training process. We report the results of our experiments on six major image classification datasets and show that the proposed approach significantly outperforms the baseline algorithms. Comparisons with popular data sampling techniques and cost-sensitive classifiers demonstrate the superior performance of our proposed method.

show abstract

Section: Results and Comparisonsmentioning

confidence: 99%

Cost-Sensitive Learning of Deep Feature Representations From Imbalanced Data

Khan

Hayat

Bennamoun

et al. 2018

IEEE Trans. Neural Netw. Learning Syst.

589

View full text Add to dashboard Cite

show abstract

“…As shown in Table1, our method for scene recognition with learned attributes (CMAP-A), performs competitively with [34] while using shorter feature vectors in relatively cheaper environment, and outperforms the others. Comparisons with [45] show that using the visual information acquired from attributes is more descriptive in the cluttered nature of MIT-indoor scenes.…”

Section: Attribute Learningmentioning

confidence: 99%

“…[34] [50] [27] [50,9,8,25,10,26,35,7]. In these studies weakly labeled datasets are leveraged for learning visual patches that are representative and discriminative.…”

Section: Learning Discriminative Patchesmentioning

confidence: 99%

ConceptMap: Mining Noisy Web Data for Concept Learning

Golge

Duygulu

2014

Computer Vision – ECCV 2014

View full text Add to dashboard Cite

Abstract. We attack the problem of learning concepts automatically from noisy Web image search results. The idea is based on discovering common characteristics shared among subsets of images by posing a method that is able to organise the data while eliminating irrelevant instances. We propose a novel clustering and outlier detection method, namely Concept Map (CMAP). Given an image collection returned for a concept query, CMAP provides clusters pruned from outliers. Each cluster is used to train a model representing a different characteristics of the concept. The proposed method outperforms the state-of-the-art studies on the task of learning from noisy web data for low-level attributes, as well as high level object categories. It is also competitive with the supervised methods in learning scene concepts. Moreover, results on naming faces support the generalisation capability of the CMAP framework to different domains. CMAP is capable to work at large scale with no supervision through exploiting the available sources.

show abstract

“…First, distinct power of learned parts is used to alleviate visual ambiguity. Recent work [8,24,15,16,27,17] discovered parts with specific visual concepts -that is, the learned part is expected to represent a cluster of visual objects. Second, unsupervised discovery of discriminative parts is dominating.…”

Section: Related Workmentioning

confidence: 99%