Self-Supervised Joint Learning for pCLE Image Denoising

Kun Yang,Haojie Zhang,Yufei Qiu,Tong Zhai,Zhiguo Zhang

DOI: https://doi.org/10.3390/s24092853

IF: 3.9

2024-04-30

Sensors

Abstract:Probe-based confocal laser endoscopy (pCLE) has emerged as a powerful tool for disease diagnosis, yet it faces challenges such as the formation of hexagonal patterns in images due to the inherent characteristics of fiber bundles. Recent advancements in deep learning offer promise in image denoising, but the acquisition of clean-noisy image pairs for training networks across all potential scenarios can be prohibitively costly. Few studies have explored training denoising networks on such pairs. Here, we propose an innovative self-supervised denoising method. Our approach integrates noise prediction networks, image quality assessment networks, and denoising networks in a collaborative, jointly trained manner. Compared to prior self-supervised denoising methods, our approach yields superior results on pCLE images and fluorescence microscopy images. In summary, our novel self-supervised denoising technique enhances image quality in pCLE diagnosis by leveraging the synergy of noise prediction, image quality assessment, and denoising networks, surpassing previous methods on both pCLE and fluorescence microscopy images.

engineering, electrical & electronic,instruments & instrumentation,chemistry, analytical

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to remove noise and improve image quality in probe - based laser confocal endomicroscopy (pCLE) images. The pCLE technology shows great potential in disease diagnosis. However, due to the inherent characteristics of the fiber bundle, its imaging results often show a hexagonal pattern, which will interfere with the accurate analysis of the image. Traditional deep - learning methods require a large number of pairs of clean images and noisy images for training. But in practical applications, especially in the medical field, the cost of obtaining these data is very high and difficult to achieve. In addition, the existing denoising methods are mainly for Gaussian noise, while the noise distribution in medical images is often non - uniform, which makes the traditional methods less effective in dealing with such complex situations.

To solve these problems, the paper proposes an innovative self - supervised joint - learning method for pCLE image denoising. This method integrates a noise prediction network, an image quality evaluation network and a denoising network, and conducts joint training in a collaborative manner. Specifically:

1. **Noise Prediction Network (N - Net)**: Predict the noise level map in the image through a convolutional neural network (CNN) to provide an estimate of the noise intensity.
2. **Image Quality Evaluation Network (Q - Net)**: Evaluate the quality of the reconstructed image and generate a quality score trend map for each image, thereby determining the best image restoration time.
3. **Denoising Network (D - Net)**: Combine the output of the noise prediction network and the noisy image to generate the final denoising result.

Compared with previous self - supervised denoising methods, this method has achieved better denoising effects on pCLE images and fluorescence microscope images. Through this joint training framework, the method proposed in the paper not only improves the image quality, but also enhances the generalization ability for unknown noise levels and types.