Semi-supervised Multi-task Learning with Chest X-Ray Images

Abstract

Discriminative models that require full supervision are inefficacious in the medical imaging domain when large labeled datasets are unavailable. By contrast, generative modeling-i.e., learning data generation and classification-facilitates semi-supervised training with limited labeled data. Moreover, generative modeling can be advantageous in accomplishing multiple objectives for better generalization. We propose a novel multi-task learning model for jointly learning a classifier and a segmentor, from chest X-ray images, through semi-supervised learning. In addition, we propose a new loss function that combines absolute KL divergence with Tversky loss (KLTV) to yield faster convergence and better segmentation performance. Based on our experimental results using a novel segmentation model, an Adversarial Pyramid Progressive Attention U-Net (APPAU-Net), we hypothesize that KLTV can be more effective for generalizing multi-tasking models while being competitive in segmentation-only tasks.

Generalized Multi-Task Learning from Substantially Unlabeled Multi-Source Medical Image Data

“…With an augmented literature review, a more detailed explanation of the methods, model architecture, and training algorithm, further details about the datasets, saliency map visualizations from multiple datasets, and additional results and discussion supported by quantitative (performance metrics tables) and qualitative (mask predictions, Bland Altman plots, ROC curves, consistency plots) characteristics. that a multi-tasking model usually outperforms its single-task counterparts (Imran and Terzopoulos, 2019;Imran, 2020). The shared encoder in the MultiMix model learns features useful for addressing both the classification and segmentation tasks.…”

Self-Supervised, Semi-Supervised, Multi-Context Learning for the Combined Classification and Segmentation of Medical Images (Student Abstract)

“…All the images are normalized and resized to 128 × 128 × 1 before feeding them to the models. We use a U-Net-like encoder-decoder network with skip connections as the segmentation mask generator and another convolutional network as the class discriminator (Imran and Terzopoulos 2019b).…”

X-Rays and Computed Tomography Scan Imaging: Instrumentation and Medical Applications

“…Technology advances of electronics made use of X-ray for digital imaging by replacing the traditional X-ray-sensitive film by electronic sensors [7][8][9][10]. Today, both convention and digital X-ray imaging modalities are the prompt and main diagnostic tools for investigating and screening the chest for viral and bacterial pneumonia, tuberculosis, lung cancer [11][12][13][14][15][16][17][18][19], enlarged heart, and blocked blood vessels [20][21][22][23][24]; the bones and teeth for fractures and infections, arthritis, bone cancer, and dental decay [25][26][27][28][29][30]; the abdomen for digestive tract problems and looking for swallowed items [31]. Moreover, other modalities for Xray imaging have been developed such as digital mammography for breast cancer screening [32].…”