FASON: First and Second Order Information Fusion Network for Texture Recognition

Dai, Xiyang; Ng, Joe Yue-Hei; Davis, Larry S.

doi:10.1109/cvpr.2017.646

Cited by 66 publications

(37 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Compact BCNN [63] Adopting Random Maclaurin Projection or Tensor Sketch Projection to reduce the dimensionality of bilinear features (e.g. from 262144 (512 2 ) to 8192); Maintain similar performance to BCNN; • FASON [46] Combining the ideas of TCNN [9] and Compact BCNN [63]. • NetVLAD [10] Plugging a VLAD like layer in a CNN network at the last CONV layer.…”

Section: Using Pretrained Generic Cnn Modelsmentioning

confidence: 99%

From BoW to CNN: Two Decades of Texture Representation for Texture Classification

et al. 2018

View full text Add to dashboard Cite

Texture is a fundamental characteristic of many types of images, and texture representation is one of the essential and challenging problems in computer vision and pattern recognition which has attracted extensive research attention over several decades. Since 2000, texture representations based on Bag of Words (BoW) and on Convolutional Neural Networks (CNNs) have been extensively studied with impressive performance. Given this period of remarkable evolution, this paper aims to present a comprehensive survey of advances in texture representation over the last two decades. More than 250 major publications are cited in this survey covering different aspects of the research, including benchmark datasets and state of the art results. In retrospect of what has been achieved so far, the survey discusses open challenges and directions for future research.

show abstract

Section: Using Pretrained Generic Cnn Modelsmentioning

confidence: 99%

From BoW to CNN: Two Decades of Texture Representation for Texture Classification

et al. 2018

View full text Add to dashboard Cite

show abstract

“…Context is important in computer vision tasks [36,24]. Temporal contextual information has been shown to be critical for temporal activity detection [9,7].…”

Section: Local and Global Temporal Contextsmentioning

confidence: 99%

Dynamic Temporal Pyramid Network: A Closer Look at Multi-scale Modeling for Activity Detection

Zhang

Dai

Wang

2019

Computer Vision – ACCV 2018

Self Cite

View full text Add to dashboard Cite

Recognizing instances at varying scales simultaneously is a fundamental challenge in visual detection problems. While spatial multiscale modeling has been well studied in object detection, how to effectively apply a multi-scale architecture to temporal models for activity detection is still under-explored. In this paper, we identify three unique challenges that need to be specifically handled for temporal activity detection. To address all these issues, we propose Dynamic Temporal Pyramid Network (DTPN), a new activity detection framework with a multiscale pyramidal architecture featuring three novel designs: (1) We sample frame sequence dynamically with different frame per seconds (FPS) to construct a natural pyramidal representation for arbitrary-length input videos. (2) We design a two-branch multi-scale temporal feature hierarchy to deal with the inherent temporal scale variation of activity instances. (3) We further exploit the temporal context of activities by appropriately fusing multi-scale feature maps, and demonstrate that both local and global temporal contexts are important. By combining all these components into a uniform network, we end up with a singleshot activity detector involving single-pass inferencing and end-to-end training. Extensive experiments show that the proposed DTPN achieves state-of-the-art performance on the challenging ActvityNet dataset.

show abstract

“…Note that (i) matrix power normalization is impractical for ResNet101 features, (ii) it cannot be computed by sketching due to Proposition 6. We also outperform FASON [10] Results on FMD. Table 3.…”

Section: Democratic Pooling With Tensor Sketchingmentioning

confidence: 87%

Second-Order Democratic Aggregation

Lin

Maji

Koniusz³

2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Aggregated second-order features extracted from deep convolutional networks have been shown to be effective for texture generation, fine-grained recognition, material classification, and scene understanding. In this paper, we study a class of orderless aggregation functions designed to minimize interference or equalize contributions in the context of second-order features and we show that they can be computed just as efficiently as their first-order counterparts and they have favorable properties over aggregation by summation. Another line of work has shown that matrix power normalization after aggregation can significantly improve the generalization of second-order representations. We show that matrix power normalization implicitly equalizes contributions during aggregation thus establishing a connection between matrix normalization techniques and prior work on minimizing interference. Based on the analysis we present γ-democratic aggregators that interpolate between sum (γ=1) and democratic pooling (γ=0) outperforming both on several classification tasks. Moreover, unlike power normalization, the γdemocratic aggregations can be computed in a low dimensional space by sketching that allows the use of very high-dimensional second-order features. This results in a state-of-the-art performance on several datasets.

show abstract

FASON: First and Second Order Information Fusion Network for Texture Recognition

Cited by 66 publications

References 23 publications

From BoW to CNN: Two Decades of Texture Representation for Texture Classification

From BoW to CNN: Two Decades of Texture Representation for Texture Classification

Dynamic Temporal Pyramid Network: A Closer Look at Multi-scale Modeling for Activity Detection

Second-Order Democratic Aggregation

Contact Info

Product

Resources

About