Background: Blood urea nitrogen to albumin ratio (BAR) has been implicated in predicting outcomes of various inflammatory-related diseases. However, the predictive value of BAR in long-term mortality in patients with acute myocardial infarction (AMI) has not yet been evaluated. Methods: In this retrospective cohort study, the patients were recruited from the Medical Information Mart for Intensive Care III (MIMIC III) database and categorized into two groups by a cutoff value of BAR. Kaplan–Meier (K-M) analysis and Cox proportional hazard model were performed to determine the predictive value of BAR in long-term mortality following AMI. In order to adjust the baseline differences, a 1:1 propensity score matching (PSM) was carried out and the results were further validated. Results: A total of 1827 eligible patients were enrolled. The optimal cutoff value of BAR for four-year mortality was 7.83 mg/g. Patients in the high BAR group tended to have a longer intensive care unit (ICU) stay and a higher rate of one-, two-, three- and four-year mortality (all p< 0.001) compared with those in the low BAR group. K-M curves indicated a significant difference in four-year survival (p< 0.001) between low and high BAR groups. The Cox proportional hazards model showed that higher BAR (> 7.83) was independently associated with increased four-year mortality in the entire cohort, with a hazard ratio (HR) of 1.478 [95% CI (1.254– 1.740), p< 0.001]. After PSM, the baseline characteristics of 312 pairs of patients in the high and low BAR groups were well balanced, and similar results were observed in K-M curve (p=0.003). Conclusion: A higher BAR (> 7.83) was associated with four-year mortality in patients with AMI. As an easily available biomarker, BAR can predict the long-term mortality in AMI patients independently.

Keypoint detection and description play a pivotal role in various robotics and autonomous applications, including Visual Odometry (VO), visual navigation, and Simultaneous Localization And Mapping (SLAM). While a myriad of keypoint detectors and descriptors have been extensively studied in conventional camera images, the effectiveness of these techniques in the context of LiDAR-generated images, i.e., reflectivity and ranges images, has not been assessed. These images have gained attention due to their resilience in adverse conditions, such as rain or fog. Additionally, they contain significant textural information that supplements the geometric information provided by LiDAR point clouds in the point cloud registration phase, especially when reliant solely on LiDAR sensors. This addresses the challenge of drift encountered in LiDAR Odometry (LO) within geometrically identical scenarios or where not all the raw point cloud is informative and may even be misleading. This paper aims to analyze the applicability of conventional image keypoint extractors and descriptors on LiDAR-generated images via a comprehensive quantitative investigation. Moreover, we propose a novel approach to enhance the robustness and reliability of LO. After extracting keypoints, we proceed to downsample the point cloud, subsequently integrating it into the point cloud registration phase for the purpose of odometry estimation. Our experiment demonstrates that the proposed approach has comparable accuracy but reduced computational overhead, higher odometry publishing rate, and even superior performance in scenarios prone to drift by using the raw point cloud. This, in turn, lays a foundation for subsequent investigations into the integration of LiDAR-generated images with LO.

Keypoint detection and description play a pivotal role in various robotics and autonomous applications including visual odometry (VO), visual navigation, and Simultaneous localization and mapping (SLAM). While a myriad of keypoint detectors and descriptors have been extensively studied in conventional camera images, the effectiveness of these techniques in the context of LiDAR-generated images, i.e. reflectivity and ranges images, has not been assessed. These images have gained attention due to their resilience in adverse conditions such as rain or fog. Additionally, they contain significant textural information that supplements the geometric information provided by LiDAR point clouds in the point cloud registration phase, especially when reliant solely on LiDAR sensors. This addresses the challenge of drift encountered in LiDAR Odometry (LO) within geometrically identical scenarios or where not all the raw point cloud is informative and may even be misleading. This paper aims to analyze the applicability of conventional image key point extractors and descriptors on LiDAR-generated images via a comprehensive quantitative investigation. Moreover, we propose a novel approach to enhance the robustness and reliability of LO. After extracting key points, we proceed to downsample the point cloud, subsequently integrating it into the point cloud registration phase for the purpose of odometry estimation. Our experiment demonstrates that the proposed approach has comparable accuracy but reduced computational overhead, higher odometry publishing rate, and even superior performance in scenarios prone to drift by using the raw point cloud. This, in turn, lays a foundation for subsequent investigations into the integration of LiDAR-generated images with LO. Our code is available on GitHub: https://github.com/TIERS/ws-lidar-as-camera-odom.