Low-dose computed tomography (LDCT) lowers risks linked to radiation exposure, but relies on advanced denoising algorithms to maintain diagnostic image quality. Reigning learning-based models aim to separate noise from tissue signals by projecting LDCT images through multiple network stages that extract latent feature maps. Naturally, separation fidelity depends on the model's ability to capture short- to long-range contextual dependencies across spatial and channel dimensions of these maps. Existing convolutional and transformer models either lack sensitivity to long-range context or suffer from efficiency-related trade-offs, limiting their effectiveness. To achieve high-fidelity LDCT denoising, here we introduce a novel denoising method, DenoMamba, that performs state-space modeling (SSM) to efficiently capture both short- and long-range context in CT images. DenoMamba leverages a novel cascaded architecture equipped with spatial SSM modules to encode spatial context and channel SSM modules comprising a gated convolution network to encode content-aware features of channel context. Contextual feature maps are then consolidated with low-level spatial features via a convolution fusion module (CFM). Comprehensive experiments at 25% and 10% dose reduction demonstrate that DenoMamba outperforms state-of-the-art convolutional, transformer and SSM denoisers with average improvements of 1.6 dB PSNR and 1.7% SSIM in image quality.

Learning-based methods offer performance leaps over traditional methods in classification analysis of high-dimensional functional MRI (fMRI) data. In this domain, deep-learning models that analyze functional connectivity (FC) features among brain regions have been particularly promising. However, many existing models receive as input temporally static FC features that summarize inter-regional interactions across an entire scan, reducing the temporal sensitivity of classifiers by limiting their ability to leverage information on dynamic FC features of brain activity. To improve the performance of baseline classification models without compromising efficiency, here we propose a novel plug-in based on a graph neural network, GraphCorr, to provide enhanced input features to baseline models. The proposed plug-in computes a set of latent FC features with enhanced temporal information while maintaining comparable dimensionality to static features. Taking brain regions as nodes and blood-oxygen-level-dependent (BOLD) signals as node inputs, GraphCorr leverages a node embedder module based on a transformer encoder to capture dynamic latent representations of BOLD signals. GraphCorr also leverages a lag filter module to account for delayed interactions across nodes by learning correlational features of windowed BOLD signals across time delays. These two feature groups are then fused via a message passing algorithm executed on the formulated graph. Comprehensive demonstrations on three public datasets indicate improved classification performance for several state-of-the-art graph and convolutional baseline models when they are augmented with GraphCorr.

Imputation of missing images via source-to-target modality translation can improve diversity in medical imaging protocols. A pervasive approach for synthesizing target images involves one-shot mapping through generative adversarial networks (GAN). Yet, GAN models that implicitly characterize the image distribution can suffer from limited sample fidelity. Here, we propose a novel method based on adversarial diffusion modeling, SynDiff, for improved performance in medical image translation. To capture a direct correlate of the image distribution, SynDiff leverages a conditional diffusion process that progressively maps noise and source images onto the target image. For fast and accurate image sampling during inference, large diffusion steps are taken with adversarial projections in the reverse diffusion direction. To enable training on unpaired datasets, a cycle-consistent architecture is devised with coupled diffusive and non-diffusive modules that bilaterally translate between two modalities. Extensive assessments are reported on the utility of SynDiff against competing GAN and diffusion models in multi-contrast MRI and MRI-CT translation. Our demonstrations indicate that SynDiff offers quantitatively and qualitatively superior performance against competing baselines.

Melanoma is a fatal skin cancer that is curable and has dramatically increasing survival rate when diagnosed at early stages. Learning-based methods hold significant promise for the detection of melanoma from dermoscopic images. However, since melanoma is a rare disease, existing databases of skin lesions predominantly contain highly imbalanced numbers of benign versus malignant samples. In turn, this imbalance introduces substantial bias in classification models due to the statistical dominance of the majority class. To address this issue, we introduce a deep clustering approach based on the latent-space embedding of dermoscopic images. Clustering is achieved using a novel center-oriented margin-free triplet loss (COM-Triplet) enforced on image embeddings from a convolutional neural network backbone. The proposed method aims to form maximally-separated cluster centers as opposed to minimizing classification error, so it is less sensitive to class imbalance. To avoid the need for labeled data, we further propose to implement COM-Triplet based on pseudo-labels generated by a Gaussian mixture model (GMM). Comprehensive experiments show that deep clustering with COM-Triplet loss outperforms clustering with triplet loss, and competing classifiers in both supervised and unsupervised settings.

In this study, an effective segmentation method is presented for defect detection on the glass surface. Defect detection on the glass surface is compelling and strenuous job for human eyes. Transparency and reflection properties of the glass surface reduce success of image processing algorithms using detection of the factors that unwanted and affecting quality of products such as crack, scratch, bubble. Traditional methods have limited success and long processing time in this process. Therefore, fast and effective method has been proposed. In the proposed method (BiasFeed CNN), bias input which is single number value traditional CNN algorithm is converted bias template. Bias template is used to balance the brightness level of the image. Input image and bias template convolution is applied bias input. Through the contribution from bias input, background reflections and negative effects arising from transparency are decreased. The developed method is fast as traditional CNN, because it does not cause significant changes in the structure of traditional CNN. 35 pieces of glass was tested using the algorithm. Damages in the glass surface and location of these damages were determined. Accuracy rates of inspected images are; sensivity %91, specificity %99, accuracy %98. BiasFeed CNN algorithm was tested on various images and it is more successful than traditional CNN algorithm.