Lesion region inpainting: an approach for pseudo-healthy image synthesis in intracranial infection imaging.

The synthesis of pseudo-healthy images, involving the generation of healthy counterparts for pathological images, is crucial for data augmentation, clinical disease diagnosis, and understanding pathology-induced changes. Recently, Generative Adversarial Networks (GANs) have shown substantial promise in this domain. However, the heterogeneity of intracranial infection symptoms caused by various infections complicates the model's ability to accurately differentiate between pathological and healthy regions, leading to the loss of critical information in healthy areas and impairing the precise preservation of the subject's identity. Moreover, for images with extensive lesion areas, the pseudo-healthy images generated by these methods often lack distinct organ and tissue structures. To address these challenges, we propose a three-stage method (localization, inpainting, synthesis) that achieves nearly perfect preservation of the subject's identity through precise pseudo-healthy synthesis of the lesion region and its surroundings. The process begins with a Segmentor, which identifies the lesion areas and differentiates them from healthy regions. Subsequently, a Vague-Filler fills the lesion areas to construct a healthy outline, thereby preventing structural loss in cases of extensive lesions. Finally, leveraging this healthy outline, a Generative Adversarial Network integrated with a contextual residual attention module generates a more realistic and clearer image. Our method was validated through extensive experiments across different modalities within the BraTS2021 dataset, achieving a healthiness score of 0.957. The visual quality of the generated images markedly exceeded those produced by competing methods, with enhanced capabilities in repairing large lesion areas. Further testing on the COVID-19-20 dataset showed that our model could effectively partially reconstruct images of other organs.

Self-supervised context-aware correlation filter for robust landmark tracking in liver ultrasound sequences.

OBJECTIVES: Respiratory motion-induced displacement of internal organs poses a significant challenge in image-guided radiation therapy, particularly affecting liver landmark tracking accuracy. METHODS: Addressing this concern, we propose a self-supervised method for robust landmark tracking in long liver ultrasound sequences. Our approach leverages a Siamese-based context-aware correlation filter network, trained by using the consistency loss between forward tracking and back verification. By effectively utilizing both labeled and unlabeled liver ultrasound images, our model, Siam-CCF , mitigates the impact of speckle noise and artifacts on ultrasonic image tracking by a context-aware correlation filter. Additionally, a fusion strategy for template patch feature helps the tracker to obtain rich appearance information around the point-landmark. RESULTS: Siam-CCF achieves a mean tracking error of 0.79 ± 0.83 mm at a frame rate of 118.6 fps, exhibiting a superior speed-accuracy trade-off on the public MICCAI 2015 Challenge on Liver Ultrasound Tracking (CLUST2015) 2D dataset. This performance won the 5th place on the CLUST2015 2D point-landmark tracking task. CONCLUSIONS: Extensive experiments validate the effectiveness of our proposed approach, establishing it as one of the top-performing techniques on the CLUST2015 online leaderboard at the time of this submission.

Deep learning for registration of region of interest in consecutive wireless capsule endoscopy frames.

BACKGROUND AND OBJECTIVE: Functional gastrointestinal disorders (FGIDs) are reported as worldwide gastrointestinal (GI) diseases. GI motility assessment can assist the diagnosis of patients with intestine motility dysfunction. Wireless capsule endoscopy (WCE) can acquire images in the gastrointestinal (GI) tract including the small intestine where other conventional endoscopes cannot penetrate, and WCE images can reveal GI motility. To generally analyze WCE frames, the high-precision registration of consecutive WCE frames is an absolute necessity. It is difficult and meaningless to register entire WCE frames on a pixel level due to the unpredictable and massive non-rigid deformation between consecutive frames, the low quality of imaging and the complex intestinal environment. Thus, the registration of region of interest (ROI) functioning in a feature level has more significance than entire frame registration. METHODS: In this paper we present Timecylce-WCE, an end-to-end automatic registration approach of ROIs on WCE images. The clinicians can determine a ROI by drawing a bounding box in any WCE frame to be registered. This proposed approach is based on a deep-learning model of time-consistency in recurrent-registering, skip-registering and self-registering cycle, and it is fully unsupervised without any label. We incorporate the global correlation map with the local correlation map in matching the features, and a novel overall loss function is designed to enable the convergence of the model. As the output, a thin-plate spline (TPS) transformed region in the template frame is highly aligned with the query ROI in a finer-grained level. To the best of our knowledge this is the first time that a deep-learning-based registration method is proposed for WCE imaging motion. RESULTS: To highlight the effectiveness of the proposed approach, our proposed method is compared with the existing non-deep-learning methods and tested in a validation dataset with labeled matching points. The presented method resulted in the best PCK@10 (Percentage of Correct Key-points, i.e., the predicted and the true joint is within the threshold - 10 pixels) of 66.49%. We also demonstrate that variants of design improved registration accuracy. CONCLUSIONS: From the experimental analysis, it is clear that our proposed method outperforms the other existing methods. This lays the groundwork for subsequent studies, such as GI motility assessment, and WCE image synthesis.