Are Vision xLSTM Embedded UNet More Reliable in Medical 3D Image Segmentation?

The advancement of developing efficient medical image segmentation has evolved from initial dependence on Convolutional Neural Networks (CNNs) to the present investigation of hybrid models that combine CNNs with Vision Transformers. Furthermore, there is an increasing focus on creating architectures that are both high-performing in medical image segmentation tasks and computationally efficient to be deployed on systems with limited resources. Although transformers have several advantages like capturing global dependencies in the input data, they face challenges such as high computational and memory complexity. This paper investigates the integration of CNNs and Vision Extended Long Short-Term Memory (Vision-xLSTM) models by introducing a novel approach called UVixLSTM. The Vision-xLSTM blocks captures temporal and global relationships within the patches extracted from the CNN feature maps. The convolutional feature reconstruction path upsamples the output volume from the Vision-xLSTM blocks to produce the segmentation output. Our primary objective is to propose that Vision-xLSTM forms a reliable backbone for medical image segmentation tasks, offering excellent segmentation performance and reduced computational complexity. UVixLSTM exhibits superior performance compared to state-of-the-art networks on the publicly-available Synapse dataset. Code is available at: https://github.com/duttapallabi2907/UVixLSTM

Mixed Attention with Deep Supervision for Delineation of COVID Infection in Lung CT

The COVID-19 pandemic, with its multiple variants, has placed immense pressure on the global healthcare system. An early effective screening and grading become imperative towards optimizing the limited available resources of the medical facilities. Computed tomography (CT) provides a significant non-invasive screening mechanism for COVID-19 infection. An automated segmentation of the infected volumes in lung CT is expected to significantly aid in the diagnosis and care of patients. However, an accurate demarcation of lesions remains problematic due to their irregular structure and location(s) within the lung. A novel deep learning architecture, Mixed Attention Deeply Supervised Network (MiADS-Net), is proposed for delineating the infected regions of the lung from CT images. Incorporating dilated convolutions with varying dilation rates, into a mixed attention framework, allows capture of multi-scale features towards improved segmentation of lesions having different sizes and textures. Mixed attention helps prioritise relevant feature maps to be probed, along with those regions containing crucial information within these maps. Deep supervision facilitates discovery of robust and discriminatory characteristics in the hidden layers at shallower levels, while overcoming the vanishing gradient. This is followed by estimating the severity of the disease, based on the ratio of the area of infected region in each lung with respect to its entire volume. Experimental results, on three publicly available datasets, indicate that the MiADS-Net outperforms several state-of-the-art architectures in the COVID-19 lesion segmentation task; particularly in defining structures involving complex geometries.

Full-scale Deeply Supervised Attention Network for Segmenting COVID-19 Lesions

Automated delineation of COVID-19 lesions from lung CT scans aids the diagnosis and prognosis for patients. The asymmetric shapes and positioning of the infected regions make the task extremely difficult. Capturing information at multiple scales will assist in deciphering features, at global and local levels, to encompass lesions of variable size and texture. We introduce the Full-scale Deeply Supervised Attention Network (FuDSA-Net), for efficient segmentation of corona-infected lung areas in CT images. The model considers activation responses from all levels of the encoding path, encompassing multi-scalar features acquired at different levels of the network. This helps segment target regions (lesions) of varying shape, size and contrast. Incorporation of the entire gamut of multi-scalar characteristics into the novel attention mechanism helps prioritize the selection of activation responses and locations containing useful information. Determining robust and discriminatory features along the decoder path is facilitated with deep supervision. Connections in the decoder arm are remodeled to handle the issue of vanishing gradient. As observed from the experimental results, FuDSA-Net surpasses other state-of-the-art architectures; especially, when it comes to characterizing complicated geometries of the lesions.