A Deep Regression Approach for Human Activity Recognition Under Partial Occlusion

In real-life scenarios, Human Activity Recognition (HAR) from video data is prone to occlusion of one or more body parts of the human subjects involved. Although it is common sense that the recognition of the majority of activities strongly depends on the motion of some body parts, which when occluded compromise the performance of recognition approaches, this problem is often underestimated in contemporary research works. Currently, training and evaluation is based on datasets that have been shot under laboratory (ideal) conditions, i.e. without any kind of occlusion. In this work, we propose an approach for HAR in the presence of partial occlusion, in cases wherein up to two body parts are involved. We assume that human motion is modeled using a set of 3D skeletal joints and also that occluded body parts remain occluded during the whole duration of the activity. We solve this problem using regression, performed by a novel deep Convolutional Recurrent Neural Network (CRNN). Specifically, given a partially occluded skeleton, we attempt to reconstruct the missing information regarding the motion of its occluded part(s). We evaluate our approach using four publicly available human motion datasets. Our experimental results indicate a significant increase of performance, when compared to baseline approaches, wherein networks that have been trained using only nonoccluded or both occluded and nonoccluded samples are evaluated using occluded samples. To the best of our knowledge, this is the first research work that formulates and copes with the problem of HAR under occlusion as a regression task.

Human Activity Recognition in the Presence of Occlusion

The presence of occlusion in human activity recognition (HAR) tasks hinders the performance of recognition algorithms, as it is responsible for the loss of crucial motion data. Although it is intuitive that it may occur in almost any real-life environment, it is often underestimated in most research works, which tend to rely on datasets that have been collected under ideal conditions, i.e., without any occlusion. In this work, we present an approach that aimed to deal with occlusion in an HAR task. We relied on previous work on HAR and artificially created occluded data samples, assuming that occlusion may prevent the recognition of one or two body parts. The HAR approach we used is based on a Convolutional Neural Network (CNN) that has been trained using 2D representations of 3D skeletal motion. We considered cases in which the network was trained with and without occluded samples and evaluated our approach in single-view, cross-view, and cross-subject cases and using two large scale human motion datasets. Our experimental results indicate that the proposed training strategy is able to provide a significant boost of performance in the presence of occlusion.

A Multimodal Fusion Approach for Human Activity Recognition

The problem of human activity recognition (HAR) has been increasingly attracting the efforts of the research community, having several applications. It consists of recognizing human motion and/or behavior within a given image or a video sequence, using as input raw sensor measurements. In this paper, a multimodal approach addressing the task of video-based HAR is proposed. It is based on 3D visual data that are collected using an RGB + depth camera, resulting to both raw video and 3D skeletal sequences. These data are transformed into six different 2D image representations; four of them are in the spectral domain, another is a pseudo-colored image. The aforementioned representations are based on skeletal data. The last representation is a "dynamic" image which is actually an artificially created image that summarizes RGB data of the whole video sequence, in a visually comprehensible way. In order to classify a given activity video, first, all the aforementioned 2D images are extracted and then six trained convolutional neural networks are used so as to extract visual features. The latter are fused so as to form a single feature vector and are fed into a support vector machine for classification into human activities. For evaluation purposes, a challenging motion activity recognition dataset is used, while single-view, cross-view and cross-subject experiments are performed. Moreover, the proposed approach is compared to three other state-of-the-art methods, demonstrating superior performance in most experiments.

An Optimized Cloud Computing Method for Extracting Molecular Descriptors

Extracting molecular descriptors from chemical compounds is an essential preprocessing phase for developing accurate classification models. Supervised machine learning algorithms offer the capability to detect "hidden" patterns that may exist in a large dataset of compounds, which are represented by their molecular descriptors. Assuming that molecules with similar structure tend to share similar physicochemical properties, large chemical libraries can be screened by applying similarity sourcing techniques in order to detect potential bioactive compounds against a molecular target. However, the process of generating these compound features is time-consuming. Our proposed methodology not only employs cloud computing to accelerate the process of extracting molecular descriptors but also introduces an optimized approach to utilize the computational resources in the most efficient way.

Higher Order Trust Ranking of LinkedIn Accounts with Iterative Matrix Methods

Trust is a fundamental sociotechnological mainstay of the Web today. There is substantial evidence about this since netizens implicitly or explicitly agree to trust virtually every Web service they use ranging from Web-based mail to e-commerce portals. Moreover the methodological framework for trusting individual netizens, primarily their identity and communications, has considerably progressed. Nevertheless, the core of fact checking for human generated content is still far from being substantially automated as most proposed smart algorithms capture inadequately fundamental human traits. One such case is the evaluation of the profile trustworthiness of LinkedIn members based on publicly available attributes available from the platform itself. A trusted profile may indirectly indicate a more suitable candidate since its contents can be easily verified. In this article a first order graph search mechanism for discovering LinkedIn trusted profiles based on a random walker is extended to higher order ranking based on a combination of functional and connectivity patterns. Results are derived for the same benchmark dataset and the first- and higher-order approaches are compared in terms of accuracy.

Superimposition of sequential scans to measure erosion on unpolished and curved human enamel

OBJECTIVES

To determine if superimposition of sequential scans can discriminate between different fluorides at step heights less than 5 μm on natural human enamel surfaces.

MATERIALS AND METHODS

Natural, unpolished, human enamel specimens (n = 60) were randomly assigned to one of three pre-treatment toothpaste slurries with a calcium silicate/fluoride, fluoride-only and a control. Baseline and post treatment scans, from a non-contacting profilometer with a 0.01 μm z-axis and <1 μm lateral scanning resolution were imported into superimposition software to define change in mean 3D step height and surface roughness following erosion in 0.3 % citric acid for 15 min. Statistical analysis conducted with two-way repeated measures ANOVA and post-hoc Tukey's multiple comparisons.

RESULTS

Confidence and resolution of superimposition and subtraction of repeated profiles from unpolished enamel revealed accuracy within 1-2 μm. The technique was able to discriminate between the fluorides demonstrating statistical differences in mean (SD) 3D step height (μm) of 1.96 (0.40) and 2.75 (0.49) (p=0.0024). There was a statistically significant increase in surface roughness for all groups after 15 min erosion compared to baseline. But no statistically significant difference between the interventions after 15 min erosion but there was compared to no fluoride (p=0.006).

CONCLUSION

Superimposition and subtraction of profiles could discriminate between fluoride interventions, which showed statistical differences in enamel loss differences less than 3 μm.

CLINICAL SIGNIFICANCE

This erosion model and data analysis workflow was able to distinguish differences between scans of 3 μm on unpolished enamel following the interaction with fluoride.

Evaluating denture cleanliness of patients in a regional dental hospital

Aims To evaluate the quality of clinical record keeping and determine quality of denture cleanliness; record baseline denture cleanliness for 60 patients; introduce denture hygiene instruction (DHI); and then re-assess the patients for improvement and enhanced record keeping.Methods Analysis of patients' denture hygiene instruction record keeping (n = 60) was undertaken retrospectively. The Denture Cleanliness Index (DCI) was utilised to assess denture cleanliness (best score 0, worst score 4). Baseline DCI scores were taken and individual DHI was delivered. After one month, patients were reviewed and scored, with record keeping quality analysed.Results At baseline, 11.7% (n = 7) of patients had DCI scores of ≤2, improving to 93.8% (n = 45) after one month, demonstrating short-term improvement in denture cleanliness. Only 63% (n = 38) of patients had evidence of a record of DHI within their notes at baseline, improving to 100% at recall.Conclusions New patient information leaflet and clinical guidelines on denture hygiene have been written and implemented. The results of this study suggest that this may be a relatively straightforward method to achieve a short-term improvement in denture cleanliness. The implementation of a DCI score is helpful in allowing patients to improve denture hygiene and its wider use is supported.

A Semantic Framework for Harvesting Vague Enterprise Knowledge from Microposts

The advent and wide proliferation of Social Web in the recent years has promoted the concept of social interaction as an important influencing factor of the way enterprises and organizations conduct business. Among the fields influenced is that of Enterprise Knowledge Management, where the increase and maintenance of the employees' active participation in the organization's knowledge management activities is pursued through the adoption of social computing approaches. In this paper we consider a prominent and increasingly applied such approach, namely enterprise microblogging, and we propose a novel way to exploit its knowledge generation and sharing capabilities in order to effectively capture and formalize enterprise knowledge that is vague. Application and empirical evaluation of the framework indicates significant potential towards this goal.

Effects of growth hormone and insulin-like growth factor-I on radiation enteritis. a comparative study

The effect of growth hormone (GH) and insulin-like growth factor-I (IGF-I) on intestinal mucosal integrity and bacterial translocation after abdominal radiation was studied in rats. Animals were divided into the following groups: I (control), II (radiation), III (radiation plus GH) and IV (radiation plus IGF-I). Radiation (1,100 cGy) was administered on the 1st day to groups II, III and IV. GH [0.25 mg/kg body weight (BW) s.c. once daily] was administered on days 1, 2 and 3 to group III, and IGF-I (0.1 mg/kg BW s.c. twice daily) was given on days 1, 2 and 3 to group IV. On day 4, animals were sacrificed. Cultures of the mesenteric lymph nodes (MLN) and blood cultures from aorta and portal vein were performed. The number of villi per centimeter (V/cm), the villus height (Vh), mitoses per crypt (M/C) and protein and DNA contents of the mucosa were evaluated in samples from the terminal ileum. Radiation increased the number of positive MLN cultures, while treatment with GH and IGF-I reduced them significantly. V/cm, Vh, M/C, protein and DNA contents were significantly increased in all irradiated animals treated with GH and IGF-I. In conclusion both GH and IGF-I are not only capable of improving the mucosal integrity but also to reduce the bacterial translocation that follows intestinal radiation. In small doses IGF-I could reproduce the effects of GH and in some parameters the effects were more pronounced, although not statistically significant. This suggests that the effects of GH on intestine could be mediated through the local and systemic generation of IGF-I.