Background Subtraction Using Gaussian Mixture Model Enhanced by Hole Filling Algorithm (GMMHF)

Comparison and implementation of motion detection methods for sperm detection and tracking

“…But if we only want to detect moving objects (moving sperms), then we can utilize this approach. Experiments showed that GMMHF [10] can detect sperm cell. Compared to other methods, GMMHF is not the most superior method, that has 0,43 of Kappa value and 98.01% of accuracy.…”

“…This research is also using Optical Flow operation to undertake object tracking process. In 2013, Nurhadiyatna et al [10] was compared 10 (ten) different background subtraction methods, to obtain moving object in the scene. In this research, Gaussian Mixture Model that is enhanced using Hole Filling Algorithm is performed.…”

“…In this paper Gaussian Mixture Model that is enhanced by Hole Filling Algorithm (GMMHF) as our previous research is used [10]. GMM is formulated as probability of every pixel as background or foreground, with its probability density function.…”

“…So that, GMMHF will produce more compound blob as an object and not a spared blob. The experiment will compare 6 (six) motion detection methods such as: GMMHF [10], Frame Difference that used in [8], GMM by C. Stauffer and W. Eric L. Grimson [13], C. Wren et al, [14], and Pixel-Based Adaptive Segmenter by M. Hofmann et al, [15].…”

“…Commonly, sperm analysis system is divided by several steps, such as background subtraction method for motion object detection that is used in this paper. Several background subtraction methods will be compared with our previous research [10].…”