Noise reduction by frame averaging: A numerical simulation for portal imaging systems

Abstract

We have studied the usefulness of both pre- and post-ADC frame summing for the purpose of reducing the effect of quantum noise and digitization noise in portal imaging systems. The study is based on the fluorescent-screen video-camera type of system. The study predicts the not-surprising result that provided the noise level at the ADC input is sufficiently large, the overall SNR can be increased by a factor of square root of M1M2, where M1 and M2 are the number of frames summed before and after the ADC. The study also predicts, somewhat unexpectedly, that there is an operating region in which increasing M1 actually decreases the SNR in the final image. To avoid this region M1 must be less than approximately 6 x 2(2B) (1 + -delta-1)1/2/(iaccf), where B is the number of ADC bits, -delta is the mean number of optical photons detected by the video camera per detected x-ray photon, iacc is the open-field number of detected x-ray photons per accelerator pulse per pixel, and f is the patient transmission factor. An equivalent statement is that the rms noise at the input to the ADC, sigma in, must exceed approximately 0.4q where q is the quantization interval of the ADC. It is possible that some systems operate in or close to this region. A second feature of this anomalous behavior is that the final image is not necessarily improved by increasing the number M2 of post-ADC-summed frames. For example, when sigma in/q = 0.2, there is no improvement in the overall rms error for M2 > 32.(ABSTRACT TRUNCATED AT 250 WORDS)

Rapid portal imaging with a high‐efficiency, large field‐of‐view detector

“…However, if the input noise to the ADC is reduced, a situation may arise in which post-ADC frame averaging will not make the output image any closer to the object transmission distribution ͑i.e., the input to the system͒. 21 Trade-offs between pre-and post-ADC frame summing can be analyzed using the approach described by Swindell and Mosleh-Shirazi. 21 With a reduced noise level a higher bit resolution will clearly be necessary.…”

“…A numerical simulation study of frame summing in a typical camera-based EPID has been reported. 21 An expression for the noise at the input of the ADC was derived in terms of in /q where in is the SD of the signal for a given image pixel before digitization and q is the quantization interval. It was found that the error introduced by the quantization in the ADC became significant relative to the ADC input noise when in /qϽ0.4.…”

ScintSim1 : A new Monte Carlo simulation code for transport of optical photons in 2D arrays of scintillation detectors

“…With the exception of some theoretically-demonstrated anomalous behavior, frame averaging can also be utilized to improve signal-to-noise ratio. [21]…”

Improved visualisation of choroidal neovascularisation by scanning laser ophthalmoscope using image averaging

“…[6][7][8][9][10][11][12][13][14][15] A theoretical analysis shows that the improvement of the SNR in the averaged image increases with the square root of the number of images. 13 This was confirmed in a three dimensional reconstruction of an electron microscopy tomography study 14 and in magnetic resonance imaging studies. 15 Some authors have reported problems with the technique in the presence of motion artefacts.…”

“…One of the standard technique of reducing image noise is to record a rapid series of images and average the signal over the images in the series. [6][7][8][9][10][11][12][13][14][15] Swindell and Mosleh-Shirazi have analysed the theoretical improvement that can be achieved by averaging serial images to reduce SNR 13 and found that the improvement is proportional to the square root of the number of images. However, in ophthalmic imaging applications eye motions limit the number of images that can be included.…”