Edge detection plays an important role in pattern recognition. So, the quality of the detected edge should be optimum for the pattern recognition applications. Perhaps, assessing the detected edge via a no-reference EDGE quality assessment metric is a good choice.
Hence, does there exist any reliable no-reference EDGE quality assessment metrics?
Yes, there are several no-reference edge quality assessment metrics that have been proposed in the literature. Some of the commonly used metrics include:
These metrics have been shown to provide reliable measures of edge quality for various types of images and distortions. However, it is important to note that no single metric can provide a complete assessment of image quality, and it is often necessary to use multiple metrics in combination to obtain a more comprehensive evaluation.
The term "no-reference edge quality assessment metric" refers to image sharpness metrics meant for gray scale and color images. They express the acutance of the images.
Article A customized acutance metric for quality control applications in MRI
The edge detectors give a binary image as output. Often all edge pixels will be represented by one and non-edge pixels will be represented by zero. Hence, the statement "measuring the quality of detected edges" does not make any sense. There are metrics used to compare the detected edge map with ground-truth edge map. The Pratt's Figure of Merrit is one of its kind. You can look for Edge Strength Similarity Index Metric (ESSIM) also. Please feel free to contact me if you need any source codes.
Yes, there exist several reliable no-reference EDGE quality assessment metrics that can be used to evaluate the quality of EDGE signals without the need for a reference signal. These metrics are designed to evaluate the quality of an EDGE signal based on its characteristics, such as the signal-to-noise ratio (SNR), the bit error rate (BER), and the modulation index.
Some of the commonly used no-reference EDGE quality assessment metrics are:
1. Modulation Error Ratio (MER): MER is a no-reference quality metric that is used to evaluate the modulation quality of an EDGE signal. It is calculated based on the constellation diagram of the received signal and provides a measure of the signal quality in terms of the deviation from the ideal constellation.
2. Signal-to-Noise Ratio (SNR): SNR is a no-reference quality metric that is used to evaluate the quality of an EDGE signal by comparing the power of the signal to the power of the noise in the signal. It is calculated based on the energy in the signal and the energy in the noise.
3. Bit Error Rate (BER): BER is a no-reference quality metric that is used to evaluate the quality of an EDGE signal by measuring the number of bit errors that occur in the signal. It is calculated based on the number of bit errors that occur in a given time period.
4. Perceptual Evaluation of Speech Quality (PESQ): PESQ is a no-reference quality metric that is used to evaluate the quality of speech signals transmitted over an EDGE network. It is based on the comparison of the transmitted and received speech signals and provides a measure of the perceived quality of the speech signal.
These metrics can be used to evaluate the quality of EDGE signals and identify potential issues with the signal, such as poor modulation quality or high levels of noise. They can also be used to optimize the performance of EDGE networks and improve the quality of the signals transmitted over the network.
From a detection perspective, there are some metrics for describing the capacity of accurate edge representation without the need for a reference; these would include measurements of the modulation transfer function (MTF) provided by ISO12233:2014 for example (and more recent versions of ISO12233). However, these are not always appropriate, which is what we in the IEEE P2020 Working Group have been trying to address. There are also Detectivity Metrics, have a look at Article Comparison of Detectability Index and Contrast Detection Probability
- Badges
- Science topic