Self-Supervised Video Representation Learning by Uncovering Spatio-Temporal Statistics.

This paper proposes a novel pretext task to address the self-supervised video representation learning problem. Specifically, given an unlabeled video clip, we compute a series of spatio-temporal statistical summaries, such as the spatial location and dominant direction of the largest motion, the spatial location and dominant color of the largest color diversity along the temporal axis, etc. Then a neural network is built and trained to yield the statistical summaries given the video frames as inputs. In order to alleviate the learning difficulty, we employ several spatial partitioning patterns to encode rough spatial locations instead of exact spatial Cartesian coordinates. Our approach is inspired by the observation that human visual system is sensitive to rapidly changing contents in the visual field, and only needs impressions about rough spatial locations to understand the visual contents. To validate the effectiveness of the proposed approach, we conduct extensive experiments with four 3D backbone networks, i.e., C3D, 3D-ResNet, R(2+1)D and S3D-G. The results show that our approach outperforms the existing approaches across these backbone networks on four downstream video analysis tasks including action recognition, video retrieval, dynamic scene recognition, and action similarity labeling. The source code is publicly available at: https://github.com/laura-wang/video_repres_sts.

View-Invariant Human Action Recognition Based on a 3D Bio-Constrained Skeleton Model.

Skeleton-based human action recognition has been a hot topic in recent years. Most existing studies are based on the skeleton data obtained from Kinect, which is noisy and unstable, in particular, in the case of occlusions. To cope with the noisy skeleton data and variation of viewpoints, this paper presents a view-invariant method for human action recognition by recovering the corrupted skeletons based on a 3D bio-constrained skeleton model and visualizing those body-level motion features obtained during the recovery process with images. The bio-constrained skeleton model is defined with two types of constraints: 1) constant bone lengths and 2) motion limits of joints. Based on the bio-constrained model, an effective method is proposed for skeleton recovery. Two types of new motion features, the Euclidean distance matrix between joints (JEDM), which contains the global structure information of the body, and the local dynamic variation of the joint Euler angles (JEAs) are used in describing human action. These two types of features are encoded into different motion images, which are fed into a two-stream convolutional neural network for learning different action patterns. The experiments on three benchmark datasets achieve better accuracy than the state-of-the-art approaches, which demonstrates the effectiveness of the proposed method.

The telomerase activity and expression of hTERT gene can serve as indicators in the anti-cancer treatment of human ovarian cancer.

OBJECTIVE: The aim of this study was to evaluate the clinicopathological significance of telomerase activity and expression of hTERT gene in human ovarian cancer. The potential value of them as indicators for chemotherapy in ovarian cancer cells was also studied. MATERIALS AND METHODS: A total of 73 samples and ovarian cancer cell lines HO-8910 and COC1 were studied. Telomerase activity was detected by PCR-TRAP-ELISA assay and the expression of the hTERT mRNA was analyzed by semi-quantitative RT-PCR. Alteration of the telomerase activity and hTERT mRNA were also analyzed in the ovarian cancer cells treated with different concentration and different time of cisplatin. Cytogenetic analysis was performed to compare the telomere status in the OH-8910 cells pre- and post-cisplatin treatment. The associations between these two markers and cisplatin induced-apoptosis were respectively analyzed in COC1 cells by the flow cytometry. RESULTS: Telomerase activity are highly increased in malignancy (0.795+/-0.168(A450-655 nm)) than borderline (0.389+/-0.174(A450-655 nm)), benign tumors (0.236+/-0.102(A450-655 nm)) and normal ovary (0.213+/-0.070(A450-655 nm)) (p < 0.05). Twenty samples showed detectable levels of hTERT. The hTERT gene positive lesion showed significantly higher telomerase activity than negative (p = 0.004). There is a significant correlation between the telomerase activity and expression of hTERT (r = 0.921). Both telomerase activity and expression of hTERT can reflect the chemotherapeutic effect of cisplatin in a time-dependent and dose-dependent manner. Treatment with 10 microM cisplatin, the hTERT mRNA decreased after 12h and completely disappeared after 48 h, whereas the telomerase activity did not decrease until 24h. Results from cytogenetic analysis and flow cytometry assay confirmed that the alterations of these two markers are associated with the anti-cancer treatment of cisplatin. CONCLUSION: Expression of hTERT gene is rate-limiting with the activation of telomerase. Both of they may be useful in the predicting of chemotherapeutic effect in ovarian cancer.