Regularized Max Pooling for Image Categorization

By Minh Hoai. 2014

56 citations recorded by the source corpus.

We propose Regularized Max Pooling (RMP) for image classification. RMP classifiesan image (oran image region) by extracting featurevectors at multiple subwindows at multiple locations and scales. Unlike Spatial Pyramid Matching where the subwindows are defined purely based on geometric correspondence, RMP accounts for the deformation of discriminative parts. The amount of deformation and the discriminative ability for multiple parts are jointly learned during training. RMP outperforms the state-of-the-art performance by a wide margin on the challenging PASCAL VOC2012 dataset for human action recognition on still images.

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Contextual Action Recognition with R*CNN (2015, methodology): “Hoai [15] uses an 8-layer network to extract fc7 features from regions at multiple locations and scales.”
Recognizing Actions Through Action-Specific Person Detection (2015, methodology): “The method of [18] based on regularized max pooling (RMP) of feature vectors at multiple scale and windows obtains a MAP of 76.”
Action Recognition in Still Images With Minimum Annotation Efforts (2016, methodology): “existing methods when they are using human bounding-boxes (both train BB and test BB) [8], [14], [16].”
Actions and Attributes from Wholes and Parts (2014, background): “[21], Hoai [16] and Simonyan and Zisserman [23], three CNN-based approaches on the task.”