Deep Learning and its Applications using Python
by Niha Kamal Basha, Surbhi Bhatia Khan, Abhishek Kumar, Arwa Mashat
9Enhanced Convolutional Neural Network
9.1 Introduction
In this chapter, a detailed discussion on enhanced convolutional neural network structure and its application in bio-signal processing have been depicted. As per that a detailed discussion on Absence Seizure Patten Detection using C-GRU-SVM model have been discussed with input as signal are shown along with proofs. Signal processing is about an analysis, synthesis, and modification of signal and is defined as “information about the attributes or behavior of some phenomenon” [1]. For example, detect pattern or informative data in the measured signal. Bio-signal processing is one of the application fields of signal processing. Some of them are Electroencephalography (EEG)—determines the brain activity, Electrocardiogram(ECG)—measures heart activity, Electromyogram (EMG)—determines specific muscle activity, and Electromyogram (EOG)—measures eye movement. The bandwidth and amplitude of each signal [2] varies, and their measurements are depicted in Table 9.1.
All these signals are of very low amplitude. Because of this instrumentation amplifiers are used to obtain a consistent signal with less noise. These signals are contaminated with unwanted signal component. Processing of these signals is tedious for extracting informative data. The general methodology for bio-signal processing to transfer raw data into clinically useful information is shown in Figure 9.1.
To extract clinically useful information, lots of difficulties ...