Training Deformable Part Models with Decorrelated Features

40 citations recorded by the source corpus.

How researchers see this paper

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

What Makes for Effective Detection Proposals? (2015, methodology): “-0.3 -7.4 0.1 -4.1 -10.2 -0.5 -7.2 3.0 -4.8 -1.7 2.5 2.0 -1.9 -2.5 -4.6 -1.5 -0.8 0.7 -0.3 -0.8 -2.0 Table 2: LM-LLDA detection results on PASCAL 2007 (with bounding box regression).”
How good are detection proposals, really? (2014, methodology): “This section analyses the detection proposals in the context of a DPM object detector, namely the LM-LLDA variant [14].”
Deformable Part Models with CNN Features (2014, methodology): “To achieve efficiency during training we exploit the LDA-based acceleration to DPM training of [5], using a whitened feature space constructed for CNN features; this reduces the computation time typically by a factor of four.”
Using k-Poselets for Detecting People and Localizing Their Keypoints (2014, methodology): “It’s also notable that we trained each k-poselet detector using a small set of only 200 negative images in order to accelerate training, as suggested in [16].”