Detection by detections non parametric detector adaptation for a video
This presentation is the property of its rightful owner.
Sponsored Links
1 / 22

Detection by Detections: Non-parametric Detector Adaptation for a Video PowerPoint PPT Presentation


  • 91 Views
  • Uploaded on
  • Presentation posted in: General

Detection by Detections: Non-parametric Detector Adaptation for a Video. Outline. Introduction Non-parametric detector adaption Binary codes with a vocabulary tree Similarity measure of the binary codes Transfer classification Identity grouping of detections Experiment Conclusion.

Download Presentation

Detection by Detections: Non-parametric Detector Adaptation for a Video

An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -

Presentation Transcript


Detection by detections non parametric detector adaptation for a video

Detection by Detections: Non-parametric Detector Adaptation for a Video


Outline

Outline

  • Introduction

  • Non-parametric detector adaption

    • Binary codes with a vocabulary tree

    • Similarity measure of the binary codes

    • Transfer classification

    • Identity grouping of detections

  • Experiment

  • Conclusion


Introduction

Introduction


Introduction1

Introduction

Extend any static-image-based object detector to video object detection.

Needs neither the original training data, nor manually labeled online examples.


Non parametric detector adaption

Non-parametric detector adaption

1.Set generic human detector work on high recall low precision point

2.To build a vocabulary tree using hierarchical k-means[19]

3.Encode the set of most confident visual detections as a set of binary vectors

[19] D. Nister and H. Stewenius. Scalable recognition with a vocabulary tree. In CVPR, 2006.


Binary codes with a vocabulary tree

Binary codes with a vocabulary tree


Binary codes with a vocabulary tree1

Binary codes with a vocabulary tree

Define a mapping T


Similarity measure of the binary codes

Similarity measure of the binary codes

||c||: the number of non-zero bits in the binary vector c


Transfer classification

Transfer classification

compose positive pool

change

representation

similarity scoring

final classification

decision


Identity grouping of detections

Identity grouping of detections

Group ID g(u) of any other example u


Experiment

Experiment

The normalization scheme evaluation


Experiment1

Experiment

The evaluation on the threshold for positive pool


Experiment2

Experiment

The tree depth exploration


Experiment3

Experiment

Performance of video object detection

Data is from the EC Funded CAVIAR project/IST 2001 37540

http://homepages.inf.ed.ac.uk/rbf/CAVIAR/


Experiment4

Experiment


Experiment5

Experiment

Present the adaption performance based on [6]

[6] P. Felzenszwalb, D. McAllester, and D. Ramanan. A discriminatively

trained, multiscale, deformable part model. In

CVPR, 2008.


Experiment6

Experiment


Experiment7

Experiment

Identity grouping

F(mode(gi)) counts the number of examples in group i with identity mode(gi)


Experiment8

Experiment

On positive pool

On all detection


Experiment9

Experiment

Difficult to be detected objects in video


Experiment10

Experiment

First row by our approach

Second row by k-means clustering


Conclusion

Conclusion

Simple and effective solution to improve the pure detection accuracy of off-shelf detectors trained from static images on target videos.


  • Login