Multistage Optimization Using a Modified Gaussian Mixture Model in Sperm Motility Tracking

Table 2

Summary of the previous methods in multiple sperm tracking.

Techniques	Advantages	Disadvantages
CASA; [28, 29]	(i) Detect the sperm’s features	(i) Requires a big sample; (ii) Fails in sperm collisions and proximity
JPDAF; [20, 21, 22]	(i) Adaptive learning; (ii) Works well with sperm collisions	(i) Computational complexity; (ii) Slow process termination
HGDT; [23]	(i) Combination of object segmentation and tracking; (ii) Self-learning capability	(i) Fails in multiple sperm tracking.
CSR-DCF; [30]	(i) Implementing of training and testing data for tracking and features extraction; (ii) Usage of sperm’s features in tracking	(i) Requires large training data; (ii) High computational complexity; (iii) Difficult to implement
ADT; [26, 31]	(i) Fast execution process; (ii) Multisperm tracking	(i) Its efficiency decreases in a high dense medium.; (ii) It works poorly when many sperms accumulate in a small space
DAT; [27]	(i) Easy to implement; (ii) Less mathematical complication	(i) Does not predict the dead and newborn sperms; (ii) It does not require a testing data (frames)