ABSTRAT:
The electrocardiogram’s (ECG) cyclic activity provides information about a person’s emotional, behaviour- al, and cardiovascular health. Noise that occurs during ac- quisition and symptomatic patterns produced by pathologies are two examples of irregular behaviours that affect the analysis of these signals. This paper presents a Deep Neural Network algorithm that learns the normal behaviour of an ECG when identifying irregular events, which is studied in two different settings: noise detection and symptomatic events triggered by multiple pathologies. Two noise detection algorithms were developed using an auto-encoder and Con- volution Neural Networks (CNN), with the binary class model achieving 98.18 percent accuracy and the multi-class model achieving 70.74 percent accuracy in distinguishing be- tween base wandering, muscle artefact, and electrode motion noise. Recurrent Neural Networks and an autoencoder with Gated Recurrent Units (GRU) configuration were used to create the arrhythmia detection algorithm. With a 56.85 percent accuracy and a 61.13 percent overall sensitivity for a 7-class model. It was determined that the machine’s learning mechanism learned characteristics of a regular ECG signal, sacrificing precision for greater generalisation at the moment. In the ECG, the frequency of sporadic events is more discriminatory than the classification of different types of events
EXISTING SYSTEM :
It was critical to collect clean signals with the same characteristics as those mentioned above for noise detection, as well as signals that clearly indicated particular forms of noise. As a result, signals from the Fantasia Database were used for both clean and noise-affected sections of regular ECG signals, which were created by adding different types of noise to said signals. The MIT-BIH Noise Stress Array, which involves wandering baseline, muscle artefact, and raw noise electrode motion, was used to generate noise-containing signals [3]. By adding these noise-corrupted signals to the Fantasia Database’s clean signals and retaining the same sampling frequency, this approach was used to mark in a controlled manner while still collecting a larger range of noise-corrupted signals. In the training process, the same volume of data of clean ECG signals with NSR was used, as well as equivalent quantities of each form of noise. As a result, each class’s percentage in the total training results was equal.
EXISTING SYSTEM DISADVANTAGES:
1.LESS ACCURACY
2. LOW EFFICIENCY
PROPOSED SYSTEM :
The proposed model for arrhythmia detection (Fig. 3.6) employs an encoder, thick layers, and an RNN with GRU architecture to form a sufficiently deep neural network. Each signal window generates a vector of dimension FD = 256 (feature dimension) after pre-processing the ECG signal, which consolidates the feature tensor N signal windows of 64 samples. After that, the tensor is transformed by a dense layer with no activation function, yielding a tensor with dimension HD = 128 (hidden dimension), each row representing a signal window. The previous tensor’s sequences of 32 rows are then fed into a GRU net, which comprises three GRU layers that move on the data sequentially along with the previous layer’s last state. The output of the last GRU layer is then passed through the dense block, which pro- duces the classification vector with size CD = 7 (classification dimension), same as the number of classes in this NN
PROPOSED SYSTEM ADVANTAGES:
1.HIGH ACCURACY
2.HIGH EFFICIENCY
SYSTEM REQUIREMENTS
SOFTWARE REQUIREMENTS:
• Programming Language : Python
• Font End Technologies : TKInter/Web(HTML,CSS,JS)
• IDE : Jupyter/Spyder/VS Code
• Operating System : Windows 08/10
HARDWARE REQUIREMENTS:
Processor : Core I3
RAM Capacity : 2 GB
Hard Disk : 250 GB
Monitor : 15″ Color
Mouse : 2 or 3 Button Mouse
Key Board : Windows 08/10
