Many slides based on P. FelzenszwalbP. Felzenszwalb General
object detection with deformable part-based models
Slide 2
Challenge: Generic object detection
Slide 3
Histograms of oriented gradients (HOG) Partition image into
blocks at multiple scales and compute histogram of gradient
orientations in each block N. Dalal and B. Triggs, Histograms of
Oriented Gradients for Human Detection, CVPR 2005Histograms of
Oriented Gradients for Human Detection 10x10 cells 20x20 cells
Image credit: N. Snavely
Slide 4
Histograms of oriented gradients (HOG) Partition image into
blocks at multiple scales and compute histogram of gradient
orientations in each block N. Dalal and B. Triggs, Histograms of
Oriented Gradients for Human Detection, CVPR 2005Histograms of
Oriented Gradients for Human Detection Image credit: N.
Snavely
Slide 5
Pedestrian detection with HOG Train a pedestrian template using
a linear support vector machine N. Dalal and B. Triggs, Histograms
of Oriented Gradients for Human Detection, CVPR 2005Histograms of
Oriented Gradients for Human Detection positive training examples
negative training examples
Slide 6
Pedestrian detection with HOG Train a pedestrian template using
a linear support vector machine At test time, convolve feature map
with template N. Dalal and B. Triggs, Histograms of Oriented
Gradients for Human Detection, CVPR 2005Histograms of Oriented
Gradients for Human Detection Template HOG feature mapDetector
response map
Slide 7
Example detections [Dalal and Triggs, CVPR 2005]
Slide 8
Are we done? Single rigid template usually not enough to
represent a category Many objects (e.g. humans) are articulated, or
have parts that can vary in configuration Many object categories
look very different from different viewpoints, or from instance to
instance Slide by N. Snavely
Slide 9
Discriminative part-based models P. Felzenszwalb, R. Girshick,
D. McAllester, D. Ramanan, Object Detection with Discriminatively
Trained Part Based Models, PAMI 32(9), 2010Object Detection with
Discriminatively Trained Part Based Models Root filter Part filters
Deformation weights
Slide 10
Discriminative part-based models P. Felzenszwalb, R. Girshick,
D. McAllester, D. Ramanan, Object Detection with Discriminatively
Trained Part Based Models, PAMI 32(9), 2010Object Detection with
Discriminatively Trained Part Based Models Multiple components
Slide 11
Discriminative part-based models P. Felzenszwalb, R. Girshick,
D. McAllester, D. Ramanan, Object Detection with Discriminatively
Trained Part Based Models, PAMI 32(9), 2010Object Detection with
Discriminatively Trained Part Based Models
Slide 12
Object hypothesis Multiscale model: the resolution of part
filters is twice the resolution of the root
Slide 13
Scoring an object hypothesis The score of a hypothesis is the
sum of filter scores minus the sum of deformation costs Filters
Subwindow features Deformation weights Displacements
Slide 14
Scoring an object hypothesis The score of a hypothesis is the
sum of filter scores minus the sum of deformation costs
Concatenation of filter and deformation weights Concatenation of
subwindow features and displacements Filters Subwindow features
Deformation weights Displacements
Slide 15
Detection Define the score of each root filter location as the
score given the best part placements:
Slide 16
Detection Define the score of each root filter location as the
score given the best part placements: Efficient computation:
generalized distance transforms For each default part location,
find the score of the best displacement Head filter Deformation
cost
Slide 17
Detection Define the score of each root filter location as the
score given the best part placements: Efficient computation:
generalized distance transforms For each default part location,
find the score of the best displacement Head filter
responsesDistance transform Head filter
Slide 18
Detection
Slide 19
Matching result
Slide 20
Training Training data consists of images with labeled bounding
boxes Need to learn the filters and deformation parameters
Slide 21
Training Our classifier has the form w are model parameters, z
are latent hypotheses Latent SVM training: Initialize w and
iterate: Fix w and find the best z for each training example
(detection) Fix z and solve for w (standard SVM training) Issue:
too many negative examples Do data mining to find hard
negatives
Slide 22
Car model Component 1 Component 2
Slide 23
Car detections
Slide 24
Person model
Slide 25
Person detections
Slide 26
Cat model
Slide 27
Cat detections
Slide 28
Bottle model
Slide 29
More detections
Slide 30
Quantitative results (PASCAL 2008) 7 systems competed in the
2008 challenge Out of 20 classes, first place in 7 classes and
second place in 8 classes BicyclesPersonBird Proposed approach