Both the matched and correlation filter is used for detection in AWGN channel. Matched filter uses a delay followed by sampler while correlation uses a multiplier followed by integrator. Both lead to optimum detection and are alternative way of implementation. Matched filter is frequently used in radars and sonar applications. DL can be used to learn to perform same function for detection under noisy channel to optimize the SNR of signal. Please see Chap 3 of Digital Communication by Sklar Bernard.

I agree with Rajnish Kumar's answer : the optimal detector in AWGN channel corresponds to a bank of correlators producing at their outputs the (log)-likelihood of each possibly transmitted waveform. It is followed by a circuit that selects the correlation channel producing the maximum likelihood at its output.

Each branch of the optimal detector can be realized :

1) either in direct-form by implementatiing a correlator ;

2) or either in an alternative form consisting in a matched filtering operation followed by a symbol rate sampler

These 2 different realizations are mathematically equivalent

Thank You !! Pascal Scalart Rajnish Kumar

Match filter and correlation filter are two different types of filters used in signal processing and image analysis.

Match filtering is a process where a signal or image is multiplied by a matched filter, which is a filter that has been designed to maximize the signal-to-noise ratio for a specific signal of interest. The output of a match filter is a measure of how well the input signal matches the reference signal. Match filtering is commonly used in radar and sonar systems.

Correlation filtering, on the other hand, is a process where a signal or image is correlated with a template or reference image. The output of a correlation filter is a measure of how well the input signal matches the reference image. Correlation filtering is commonly used in pattern recognition and image processing.

Which filter to use depends on the specific application and the nature of the input signal. In general, match filters are more effective for signals with a high signal-to-noise ratio, while correlation filters are more effective for signals with low signal-to-noise ratios.

Deep neural networks (DNNs) have shown great potential in signal processing and image analysis tasks, including pattern recognition and object detection. DNNs can be used to learn the relationship between input signals and the desired output, allowing for more accurate and robust signal processing. DNNs can be used for both match filtering and correlation filtering, as well as many other applications in signal processing and image analysis.