No abstract

This paper introduces the Masked Voxel Jigsaw and Reconstruction (MV-JAR) method for LiDAR-based selfsupervised pre-training and a carefully designed dataefficient 3D object detection benchmark on the Waymo dataset. Inspired by the scene-voxel-point hierarchy in downstream 3D object detectors, we design masking and reconstruction strategies accounting for voxel distributions in the scene and local point distributions within the voxel. We employ a Reversed-Furthest-Voxel-Sampling strategy to address the uneven distribution of LiDAR points and propose MV-JAR, which combines two techniques for modeling the aforementioned distributions, resulting in superior performance. Our experiments reveal limitations in previous dataefficient experiments, which uniformly sample fine-tuning splits with varying data proportions from each LiDAR sequence, leading to similar data diversity across splits. To address this, we propose a new benchmark that samples scene sequences for diverse fine-tuning splits, ensuring adequate model convergence and providing a more accurate evaluation of pre-training methods. Experiments on our Waymo benchmark and the KITTI dataset demonstrate that MV-JAR consistently and significantly improves 3D detection performance across various data scales, achieving up to a 6.3% increase in mAPH compared to training from scratch. Codes and the benchmark will be available at https://github.com/SmartBot-PJLab/MV-JAR.

Backgroud: Cancer is a major hazard to human health. Recently, small nucleolar RNA (snoRNA) has been found to be involved in the occurrence and development of cancer, which has potential diagnostic, prognostic and therapeutic value. The purpose of this study is to use the bibliometrics method to sort out and study the previous published papers. Methods We collected articles from the Web of Science Core Collection database in the field of snoRNA and cancer. Then, we used VOSviewer, Citespace, WPS and other software to visualize authors, Finally, we interpreted the data and analyzed the hotspots and frontiers of the research. Results The number of articles in this field was low in the early period, but exploded since 2008. According to the calculation of Prince's law, we believed that a stable cooperative group had been formed in this field. Chu, Liang and Montanaro, Lorenzo published the most papers, while Jiang, Feng were cited the most times. Three institutions published the most articles, namely Wuhan Univ, China Med Univ and Guangxi Med Univ. The journal with the most articles was Oncotarget. Through the analysis of countries/regions, it was found that the country with the most published articles was China. The analysis of keywords and burst words indicated that early studies mainly focused on the molecular mechanisms, but in recent years, it has gradually shifted to the direction of diagnosis, prognosis and therapy. Conclusion The research of snoRNA and cancer was a hot topic in recent years. Through analysis, we found that snoRNA was involved in the molecular mechanism of cancer development and can be used as a biomarker for clinical diagnosis and prognosis.