Related Experiment Video
Updated: Jun 27, 2025
11:33
Using Tomoauto: A Protocol for High-throughput Automated Cryo-electron Tomography
Published on: January 30, 2016
10.9K
Transformer enhanced autoencoder rendering cleaning of noisy optical coherence tomography images
Hanya Ahmed1, Qianni Zhang1, Robert Donnan2
1Queen Mary University of London, School of Electronic Engineering and Computer Science, London, United Kingdom.
Journal of Medical Imaging (Bellingham, Wash.)
|May 2, 2024
Summary
A new deep-learning framework, Transformer Enhanced Autoencoder Rendering (TEAR), effectively removes noise from Optical Coherence Tomography (OCT) images. This advanced denoising method significantly improves image quality for better medical diagnoses.
Area of Science:
- Medical Imaging
- Artificial Intelligence
Background:
- Optical Coherence Tomography (OCT) is crucial for medical imaging, particularly in ophthalmology.
- Image noise in OCT scans impedes accurate diagnosis and detailed analysis.
Purpose of the Study:
- To introduce a novel deep-learning framework for denoising OCT images.
- To enhance the visual quality and diagnostic utility of OCT scans.
Main Methods:
- A region-based, deep-learning denoising framework utilizing a hybrid model named Transformer Enhanced Autoencoder Rendering (TEAR).
- Implementation of attention gates within TEAR to focus on foreground denoising and background removal.
- Adaptive cleaning of various noise artifacts common in OCT images.
Main Results:
- TEAR demonstrated significant improvements in key performance metrics across dental and retinal OCT datasets.
- Achieved a peak signal-to-noise ratio (PSNR) of 27.9 dB and SSIM of 0.9 for dental data.
- For retinal data, metrics included PSNR of 24.6 dB and SSIM of 0.64.
Conclusions:
- The proposed TEAR framework effectively removes speckle noise from OCT images.
- TEAR outperforms traditional and existing deep-learning denoising algorithms in image quality enhancement.
Related Concept Videos
01:25Transformers with Off-Nominal Turns Ratios
151
In scenarios involving parallel transformers with disparate ratings, developing per-unit models requires accommodating off-nominal turns ratios. This situation arises when the selected base voltages are not proportional to the transformer’s voltage ratings. Consider a transformer where the rated voltages are related by the term a. If the chosen voltage bases satisfy a relationship involving term b, term c is defined as the ratio of these bases. This ratio is then substituted into the...
151
01:21Energy Losses in Transformers
864
In an ideal transformer, it is assumed that there are no energy losses, and, hence, all the power at the primary winding is transferred to the secondary winding. However, in reality, the transformers always have some energy losses, and, hence, the output power obtained at the secondary winding is less than the input power at the primary winding due to energy losses.
There are four main reasons for energy losses in transformers.
The first cause can be the high resistance of the...
There are four main reasons for energy losses in transformers.
The first cause can be the high resistance of the...
864
01:16Types Of Transformers
971
Transformers can provide desired voltages to a circuit by modifying the number of turns in the secondary windings.
If the ratio of the number of turns in the secondary winding to that of the primary winding is greater than one, then the transformer is said to be a step-up transformer. In a step-up transformer, the voltage at the secondary winding is greater than the voltage applied at the primary winding.
However, if this ratio is less than one, the transformer is said to be a step-down...
If the ratio of the number of turns in the secondary winding to that of the primary winding is greater than one, then the transformer is said to be a step-up transformer. In a step-up transformer, the voltage at the secondary winding is greater than the voltage applied at the primary winding.
However, if this ratio is less than one, the transformer is said to be a step-down...
971
01:10Reconstruction of Signal using Interpolation
194
Signal processing techniques are essential for accurately converting continuous signals to digital formats and vice versa. When a continuous signal is sampled with a period T, the resulting sampled signal exhibits replicas of the original spectrum in the frequency domain, spaced at intervals equal to the sampling frequency. To handle this sampled signal, a zero-order hold method can be applied, which creates a piecewise constant signal by retaining each sample's value until the next...
194