It is found in this research to adopt a new classifier for diagnosing
Cardiac Arrhythmias depending on detecting the Electrocardiograph (ECG),
where the classifier can identify heart beats and extract its features. Using
these features we can deci
de if the heart beat is healthy or disordered.
Beside detection normal heart beats, the research focused on detection
two diseases:
1. Premature Ventricular Contraction PVC.
2. Premature Atrial Contraction PAC.
The new classifier diagnosed the two diseases with a very high quality
where the accuracy average is 97.56%.
The new classifier is developed depending on algorithms of ANFIS
Adaptive Neural Fuzzy Inference System. System includes two consecutive
neural networks; first one sorts the heart beats to two types: normal and
abnormal were the second diagnose the disease of the disordered heartbeats
only.
This new classifier offered higher levels of efficiency and accuracy in
the comparison with the internationally known classifiers.
Forty eight dairy cows, 5-9 years old, were used to determine
their response for using of equine Chorionic Gonadotropine (eCG)
and Cloprostenol associated with PRID on post-partum anoestrus
treatment in order to increase the fertility.
The entry of computer to many areas, such as medical field, led to develop new
technique that has led to the prosperity of these areas, and helped doctors to detect and
diagnose diseases accurately and credibility, where the experience of the docto
r in addition
to the accuracy of computer lead to access to the credibility of high patient and save
human lives.
A new approach for cardiac diseases detection and classification in ECG signals
images is proposed using Adaptive Neuro Fuzzy Inference System ANFIS.
The proposed approach is applied on database containing (147) ECG images,
each of them accompanied with its medical report. The medical reports were used to
validate the detection and classification.
The proposed method achieved a relatively high accuracy (97%) in detection and
classification processes.
The proposed approach is developed using MATLAB, and based on its libraries,
image processing, neural network and fuzzy logic.