Advancements in SLAM, Visual Impairments, and Dermatological Diagnostics

The field of SLAM and autonomous navigation is moving towards more reliable and reproducible multimodal datasets, with a focus on open-hardware designs and robust calibration and synchronization pipelines. This allows for more accurate and efficient development and evaluation of SLAM algorithms. Meanwhile, innovative applications of deep learning and computer vision are being explored for early detection of visual impairments and automated dermatological diagnostics. These advances have the potential to improve treatment effectiveness and enable accessible screenings worldwide. Noteworthy papers include: SMapper, a novel open-hardware platform for SLAM research, which establishes a robust foundation for advancing SLAM algorithm development and evaluation. The KidsVisionCheck application, which provides highly reliable performance for pediatric vision screenings using red-eye reflex images and deep neural networks.

Sources

SMapper: A Multi-Modal Data Acquisition Platform for SLAM Benchmarking

Early Detection of Visual Impairments at Home Using a Smartphone Red-Eye Reflex Test

Patch-based Automatic Rosacea Detection Using the ResNet Deep Learning Framework

Privacy-Preserving Automated Rosacea Detection Based on Medically Inspired Region of Interest Selection

Event-LAB: Towards Standardized Evaluation of Neuromorphic Localization Methods

Built with on top of