The recent developments in the research area highlight significant advancements in medical imaging, computer vision, and tactile sensing technologies. A common theme across these advancements is the integration of deep learning and machine learning techniques to solve complex problems, such as continual learning, sparse-view reconstruction, and tactile sensing. In medical imaging, there's a notable shift towards methods that can handle sequential learning and adapt to new information without forgetting previously learned knowledge, as seen in continual self-supervised learning approaches. These methods are particularly beneficial for applications like chest CT imaging, where they enable the model to learn richer and more robust feature representations. Additionally, the field is seeing innovative solutions to challenges in sparse-view CT reconstruction, with novel frameworks that leverage dual-domain deep learning and hierarchical decomposition to improve reconstruction performance. In the realm of tactile sensing, advancements are being made towards more compact, efficient, and accurate sensors capable of continuous sensing on large surfaces, addressing limitations of traditional sensors. These developments not only enhance the capabilities of tactile sensors in applications like surface reconstruction and defect detection but also open new possibilities for their use in space-restricted settings.

Noteworthy Papers

• Continual Self-supervised Learning Considering Medical Domain Knowledge in Chest CT Images: Introduces a novel continual self-supervised learning method that effectively captures the relationship between previously learned knowledge and new information, enhancing the model's ability to learn meaningful representations.
• FatesGS: Fast and Accurate Sparse-View Surface Reconstruction using Gaussian Splatting with Depth-Feature Consistency: Presents an innovative sparse-view reconstruction framework that leverages intra-view depth and multi-view feature consistency for accurate surface reconstruction, significantly outperforming state-of-the-art methods in speed and quality.
• End-to-End Deep Learning for Interior Tomography with Low-Dose X-ray CT: Proposes a novel end-to-end learning approach using dual-domain CNNs to address coupled artifacts in low-dose and interior tomography, demonstrating superior performance over conventional methods.
• GelBelt: A Vision-based Tactile Sensor for Continuous Sensing of Large Surfaces: Introduces a novel vision-based tactile sensor designed for continuous surface sensing, showing promising results in shape reconstruction and surface fusion for large-scale surfaces.
• Dynamic Prototype Rehearsal for Continual Learning in ECG Arrhythmia Detection: Develops a novel CL method for ECG arrhythmia detection that introduces dynamic prototype rehearsal memory, ensuring effective knowledge retention across sessions and outperforming state-of-the-art methods.
• ThinTact: Thin Vision-Based Tactile Sensor by Lensless Imaging: Proposes a novel lensless vision-based tactile sensor with a compact design and real-time tactile sensing capabilities, demonstrating practical applicability in diverse applications.