Skip to Main Content
 

Global Search Box

 
 
 
 

ETD Abstract Container

Abstract Header

Empirical Evaluation of AdaBoost Method in Detecting Transparent and Occluded Objects

Abstract Details

2018, Master of Computing and Information Systems, Youngstown State University, Department of Computer Science and Information Systems.
Detecting and counting nano-particles in the Transmission Electron Microscopy (TEM) images is a challenging task due to two reasons: (1) The particles are semi-transparent which means that the backgrounds and objects have similar image characteristics. As a result, it is extremely difficult to separate the positive samples and the negative samples with a single or simple image feature; (2) Particles are often severely occluded (overlapped) and hence it is impossible to select a large number of clean positive samples to train a good classifier, which in turn significantly affects the detection outcomes. In this thesis, a series of empirical experiments and data analysis were conducted to compare the performances of two popular image features: Haar feature and Local Binary Pattern (LBP), within the framework of Cascade AdaBoost algorithm. It was found that the two features exhibited complex relationships with respect to several key training parameters and performance metrics, including the training time, sample size, true positive rate, false alarm rate and detection window size, etc. The experimental results and insights gained from this study help build a solid foundation upon which more detailed investigations can be carried out in the future.
Yong Zhang, PhD (Advisor)
John Sullins, PhD (Committee Member)
Feng Yu, PhD (Committee Member)
48 p.

Recommended Citations

Citations

  • Tamang, S. (2018). Empirical Evaluation of AdaBoost Method in Detecting Transparent and Occluded Objects [Master's thesis, Youngstown State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ysu1527599823503772

    APA Style (7th edition)

  • Tamang, Sujan. Empirical Evaluation of AdaBoost Method in Detecting Transparent and Occluded Objects. 2018. Youngstown State University, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ysu1527599823503772.

    MLA Style (8th edition)

  • Tamang, Sujan. "Empirical Evaluation of AdaBoost Method in Detecting Transparent and Occluded Objects." Master's thesis, Youngstown State University, 2018. http://rave.ohiolink.edu/etdc/view?acc_num=ysu1527599823503772

    Chicago Manual of Style (17th edition)