Deep-Based Quality Assessment of Medical Images Through Domain Adaptation

Abstract

Predicting the quality of multimedia content is often needed in different fields. In some applications, quality metrics are crucial with a high impact, and can affect decision making such as diagnosis from medical multimedia. In this paper, we focus on such applications by proposing an efficient and shallow model for predicting the quality of medical images without reference from a small amount of annotated data. Our model is based on convolution self-attention that aims to model complex representation from relevant local characteristics of images, which itself slide over the image to interpolate the global quality score. We also apply domain adaptation learning in unsupervised and semi-supervised manner. The proposed model is evaluated through a dataset composed of several images and their corresponding subjective scores. The obtained results showed the efficiency of the proposed method, but also, the relevance of the applying domain adaptation to generalize over different multimedia domains regarding the downstream task of perceptual quality prediction. 1

Enhancing Immersive Experiences through 3D Point Cloud Analysis: A Novel Framework for Applying 2D Visual Saliency Models to 3D Point Clouds

“…By exploring a range of viewpoints and their influence on 2D saliency predictions, we not only validate the capability of 2D models prediction to be adapted to 3D contexts but also enable the synthesis of robust 3D saliency maps. Moreover, our framework opens avenues for developing deep learning methods that leverage projected saliency data, advocating for advanced techniques like domain adaptation [39], [40] to bridge the gap between 2D and 3D saliency predictions further. These efforts aim to refine saliency prediction tools for immersive environments, ensuring that salient features stand out, enhancing the user's visual experience.…”