RESUMEN
Efficient point cloud compression is essential for applications like virtual and mixed reality, autonomous driving, and cultural heritage. This paper proposes a deep learning-based inter-frame encoding scheme for dynamic point cloud geometry compression. We propose a lossy geometry compression scheme that predicts the latent representation of the current frame using the previous frame by employing a novel feature space inter-prediction network. The proposed network utilizes sparse convolutions with hierarchical multiscale 3D feature learning to encode the current frame using the previous frame. The proposed method introduces a novel predictor network for motion compensation in the feature domain to map the latent representation of the previous frame to the coordinates of the current frame to predict the current frame's feature embedding. The framework transmits the residual of the predicted features and the actual features by compressing them using a learned probabilistic factorized entropy model. At the receiver, the decoder hierarchically reconstructs the current frame by progressively rescaling the feature embedding. The proposed framework is compared to the state-of-the-art Video-based Point Cloud Compression (V-PCC) and Geometry-based Point Cloud Compression (G-PCC) schemes standardized by the Moving Picture Experts Group (MPEG). The proposed method achieves more than 88% BD-Rate (Bjøntegaard Delta Rate) reduction against G-PCCv20 Octree, more than 56% BD-Rate savings against G-PCCv20 Trisoup, more than 62% BD-Rate reduction against V-PCC intra-frame encoding mode, and more than 52% BD-Rate savings against V-PCC P-frame-based inter-frame encoding mode using HEVC. These significant performance gains are cross-checked and verified in the MPEG working group.
RESUMEN
BACKGROUND: There are few reports describing the epidemiology of visceral leishmaniasis (VL) in Somalia. Over the years 2002 to 2005, a yearly average of 140 patients were reported from the Huddur centre in Bakool region, whereas in 2006, this number rose to 1002 patients. Given the limited amount of information on VL and the opportunity to compare features with the studies done in 2000 in this part of Somalia, we describe the epidemiologic and clinical features of patients who presented to the Huddur treatment centre of Bakool region, Somalia, using data routinely collected over a five-year observation period (2002-2006). METHODOLOGY: Methods used included the analysis of routine data on VL cases treated in the Huddur treatment centre, a retrospective study of records of patients admitted between 2004 and 2006, community leaders interviews, and analysis of blood specimens taken for parasite species identification in Antwerp Institute of Tropical Medicine. PRINCIPAL FINDINGS: A total of 1671 VL patients were admitted to the Huddur centre from January 2002 until December 2006. Nearly all patients presented spontaneously to the health centre. Since 2002, the average patient load was stable, with an average of 140 admissions per year. By the end of 2005, the number of admissions dramatically increased to reach a 7-fold increase in 2006. The genotype of L. donovani identified in 2006 was similar to the one reported in 2002. 82% of total patients treated for VL originated from two districts of Bakool region, Huddur and Tijelow districts. Clinical recovery rate was 93.2% and case fatality rate 3.9%. CONCLUSIONS: After four years of low but constant VL case findings, a major increase in VL was observed over a 16-month period in the Huddur VL centre. The profile of the patients was pediatric and mortality relatively low. Decentralized treatment centers, targeted active screening, and community sensitization will help decrease morbidity and mortality from VL in this endemic area. The true magnitude of VL in Somalia remains unknown. Further documentation to better understand transmission dynamics and thus define appropriate control measures will depend on the stability of the context and safe access to the Somali population.