Keypoint of interest based on spatio-temporal feature considering mutual dependency and camera motion

Takahiro Suzuki, Takeshi Ikenaga

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Recently, cloud systems start to be utilized for services to analyze user's data in the region of computer vision. In these services, keypoints are extracted from images or videos and the data is identified by machine learning with large database of cloud. Conventional keypoint extraction algorithms utilize only spatial information and many unnecessary keypoints for recognition are detected. Thus, the systems have to communicate large data and require processing time of descriptor calculations. This paper proposes a spatio-temporal keypoint extraction algorithm that detects only Keypoints of Interest (KOI) based on spatio-temporal feature considering mutual dependency and camera motion. The proposed method includes an approximated Kanade-Lucas-Tomasi (KLT) tracker to calculate the positions of keypoints and optical flow. This algorithm calculates the weight at each keypoint using two kinds of features: intensity gradient and optical flow. It reduces noise of extraction by comparing with states of surrounding keypoints. The camera motion estimation is added and it calculates camera-motion invariant optical flow. Evaluation results show that the proposed algorithm achieves 95% reduction of keypoint data and 53% reduction of computational complexity comparing a conventional keypoint extraction. KOI are extracted in the region whose motion and gradient are large.

Original languageEnglish
Title of host publicationMMEDIA - International Conferences on Advances in Multimedia
PublisherInternational Academy, Research and Industry Association, IARIA
Pages20-25
Number of pages6
ISBN (Print)9781612083209
Publication statusPublished - 2014
Event6th International Conferences on Advances in Multimedia, MMEDIA 2014 - Nice
Duration: 2014 Feb 232014 Feb 27

Other

Other6th International Conferences on Advances in Multimedia, MMEDIA 2014
CityNice
Period14/2/2314/2/27

Fingerprint

Optical flows
Cameras
Motion estimation
Computer vision
Learning systems
Computational complexity
Processing

Keywords

  • Keypoint extraction
  • SIFT
  • Temporal analysis

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Vision and Pattern Recognition

Cite this

Suzuki, T., & Ikenaga, T. (2014). Keypoint of interest based on spatio-temporal feature considering mutual dependency and camera motion. In MMEDIA - International Conferences on Advances in Multimedia (pp. 20-25). International Academy, Research and Industry Association, IARIA.

Keypoint of interest based on spatio-temporal feature considering mutual dependency and camera motion. / Suzuki, Takahiro; Ikenaga, Takeshi.

MMEDIA - International Conferences on Advances in Multimedia. International Academy, Research and Industry Association, IARIA, 2014. p. 20-25.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Suzuki, T & Ikenaga, T 2014, Keypoint of interest based on spatio-temporal feature considering mutual dependency and camera motion. in MMEDIA - International Conferences on Advances in Multimedia. International Academy, Research and Industry Association, IARIA, pp. 20-25, 6th International Conferences on Advances in Multimedia, MMEDIA 2014, Nice, 14/2/23.
Suzuki T, Ikenaga T. Keypoint of interest based on spatio-temporal feature considering mutual dependency and camera motion. In MMEDIA - International Conferences on Advances in Multimedia. International Academy, Research and Industry Association, IARIA. 2014. p. 20-25
Suzuki, Takahiro ; Ikenaga, Takeshi. / Keypoint of interest based on spatio-temporal feature considering mutual dependency and camera motion. MMEDIA - International Conferences on Advances in Multimedia. International Academy, Research and Industry Association, IARIA, 2014. pp. 20-25
@inproceedings{687a77131e6f44f2be23e975bd6757ae,
title = "Keypoint of interest based on spatio-temporal feature considering mutual dependency and camera motion",
abstract = "Recently, cloud systems start to be utilized for services to analyze user's data in the region of computer vision. In these services, keypoints are extracted from images or videos and the data is identified by machine learning with large database of cloud. Conventional keypoint extraction algorithms utilize only spatial information and many unnecessary keypoints for recognition are detected. Thus, the systems have to communicate large data and require processing time of descriptor calculations. This paper proposes a spatio-temporal keypoint extraction algorithm that detects only Keypoints of Interest (KOI) based on spatio-temporal feature considering mutual dependency and camera motion. The proposed method includes an approximated Kanade-Lucas-Tomasi (KLT) tracker to calculate the positions of keypoints and optical flow. This algorithm calculates the weight at each keypoint using two kinds of features: intensity gradient and optical flow. It reduces noise of extraction by comparing with states of surrounding keypoints. The camera motion estimation is added and it calculates camera-motion invariant optical flow. Evaluation results show that the proposed algorithm achieves 95{\%} reduction of keypoint data and 53{\%} reduction of computational complexity comparing a conventional keypoint extraction. KOI are extracted in the region whose motion and gradient are large.",
keywords = "Keypoint extraction, SIFT, Temporal analysis",
author = "Takahiro Suzuki and Takeshi Ikenaga",
year = "2014",
language = "English",
isbn = "9781612083209",
pages = "20--25",
booktitle = "MMEDIA - International Conferences on Advances in Multimedia",
publisher = "International Academy, Research and Industry Association, IARIA",

}

TY - GEN

T1 - Keypoint of interest based on spatio-temporal feature considering mutual dependency and camera motion

AU - Suzuki, Takahiro

AU - Ikenaga, Takeshi

PY - 2014

Y1 - 2014

N2 - Recently, cloud systems start to be utilized for services to analyze user's data in the region of computer vision. In these services, keypoints are extracted from images or videos and the data is identified by machine learning with large database of cloud. Conventional keypoint extraction algorithms utilize only spatial information and many unnecessary keypoints for recognition are detected. Thus, the systems have to communicate large data and require processing time of descriptor calculations. This paper proposes a spatio-temporal keypoint extraction algorithm that detects only Keypoints of Interest (KOI) based on spatio-temporal feature considering mutual dependency and camera motion. The proposed method includes an approximated Kanade-Lucas-Tomasi (KLT) tracker to calculate the positions of keypoints and optical flow. This algorithm calculates the weight at each keypoint using two kinds of features: intensity gradient and optical flow. It reduces noise of extraction by comparing with states of surrounding keypoints. The camera motion estimation is added and it calculates camera-motion invariant optical flow. Evaluation results show that the proposed algorithm achieves 95% reduction of keypoint data and 53% reduction of computational complexity comparing a conventional keypoint extraction. KOI are extracted in the region whose motion and gradient are large.

AB - Recently, cloud systems start to be utilized for services to analyze user's data in the region of computer vision. In these services, keypoints are extracted from images or videos and the data is identified by machine learning with large database of cloud. Conventional keypoint extraction algorithms utilize only spatial information and many unnecessary keypoints for recognition are detected. Thus, the systems have to communicate large data and require processing time of descriptor calculations. This paper proposes a spatio-temporal keypoint extraction algorithm that detects only Keypoints of Interest (KOI) based on spatio-temporal feature considering mutual dependency and camera motion. The proposed method includes an approximated Kanade-Lucas-Tomasi (KLT) tracker to calculate the positions of keypoints and optical flow. This algorithm calculates the weight at each keypoint using two kinds of features: intensity gradient and optical flow. It reduces noise of extraction by comparing with states of surrounding keypoints. The camera motion estimation is added and it calculates camera-motion invariant optical flow. Evaluation results show that the proposed algorithm achieves 95% reduction of keypoint data and 53% reduction of computational complexity comparing a conventional keypoint extraction. KOI are extracted in the region whose motion and gradient are large.

KW - Keypoint extraction

KW - SIFT

KW - Temporal analysis

UR - http://www.scopus.com/inward/record.url?scp=84907902411&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907902411&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84907902411

SN - 9781612083209

SP - 20

EP - 25

BT - MMEDIA - International Conferences on Advances in Multimedia

PB - International Academy, Research and Industry Association, IARIA

ER -