We describe an end-to-end method for recovering 3D human body mesh from single images and monocular videos. Different from the existing methods try to obtain all the complex 3D pose, shape, and camera parameters from one coupling feature, we propose a skeleton-disentangling based framework, which divides this task into multi-level spatial and temporal granularity in a decoupling manner. In spatial, we propose an effective and pluggable "disentangling the skeleton from the details" (DSD) module. It reduces the complexity and decouples the skeleton, which lays a good foundation for temporal modeling. In temporal, the selfattention based temporal convolution network is proposed to efficiently exploit the short and long-term temporal cues. Furthermore, an unsupervised adversarial training strategy, temporal shuffles and order recovery, is designed to promote the learning of motion dynamics. The proposed method outperforms the state-of-the-art 3D human mesh recovery methods by 15.4% MPJPE and 23.8% PA-MPJPE on Human3.6M. State-of-the-art results are also achieved on the 3D pose in the wild (3DPW) dataset without any fine-tuning. Especially, ablation studies demonstrate that skeleton-disentangled representation is crucial for better temporal modeling and generalization. The code is released at https://github.com

Rationale and Objectives Magnetic resonance imaging (MRI) studies reveal that the atrophy of the corpus callosum (CC) is involved in early Alzheimer’s disease (AD). This study investigates when and how the callosal change occur in the early course of AD. Materials and Methods High-resolution structural MRI were obtained from 196 old people, subjects were characterized using the Clinical Dementia Rating (CDR), 98 healthy controls were nondemented (CDR 0), 70 patients had a clinical diagnosis of AD in very mild (CDR 0.5) and 28 patients in mild (CDR 1) dementia stage. A semi-automatic segmentation method was used to extract the CC on the midsagittal plane (MSP). We measured the total and regional areas of CC. Results Results indicate that the callosal atrophy occurs in the subjects when their CDR is 0.5. The area of genu and rostral body of CC of the healthy controls (CDR 0) is significantly different from that of the subjects with very mild dementia (CDR 0.5) (p< .05). The significant difference can also be found in the area of rostral and midbody of CC between the subjects with very mild dementia (CDR 0.5) and those with mild dementia (CDR 1) (p< .05). Conclusion The callosal atrophy can be detected in the subjects when their CDR is 0.5. The change of CC in the early stage of AD indicates an anterior-to-posterior atrophic process when the degree of dementia increases (CDR 0→0.5→1).

No abstract

Background: We investigated the rate of corpus callosum (CC) atrophy and its association with cognitive decline in early Alzheimer's disease (AD). Methods: We used publicly available longitudinal MRI data corresponding to 2 or more visits from 137 subjects characterized using the Clinical Dementia Rating (CDR) score. We classified these subjects into 3 groups according to the progression of their cognitive status: a healthy control group (CDR 0→0, n = 72), a decliner group (CDR 0→0.5, n = 14) and an AD group (CDR 0.5→0.5/1, n = 51). We measured the CC area on the midsagittal plane and calculated the atrophy rate between 2 or more visits. The correlation between the CC atrophy rate and annualized Mini Mental State Examination (MMSE) change was also analyzed. Results: The results indicated that the baseline CC area was larger in the healthy control group compared to the AD group, whereas the CC atrophy rate was higher in the AD group relative to the control and decliner groups. The CC atrophy rate was also correlated with the annualized MMSE change in AD patients (p < 0.05). Conclusion: Callosal atrophy is present even in early AD and subsequently accelerates, such that the rate of CC atrophy is associated with cognitive decline in AD patients.

No abstract