Constructing Maximum Entropy Language Models for Movie Review Subjectivity Analysis

Document subjectivity analysis has become an important aspect of web text content mining. This problem is similar to traditional text categorization, thus many related classification techniques can be adapted here. However, there is one significant difference that more language or semantic information is required for better estimating the subjectivity of a document. Therefore, in this paper, our focuses are mainly on two aspects. One is how to extract useful and meaningful language features, and the other is how to construct appropriate language models efficiently for this special task. For the first issue, we conduct a Global-Filtering and Local-Weighting strategy to select and evaluate language features in a series of n-grams with different orders and within various distance-windows. For the second issue, we adopt Maximum Entropy （MaxEnt） modeling methods to construct our language model framework. Besides the classical MaxEnt models, we have also constructed two kinds of improved models with Gaussian and exponential priors respectively. Detailed experiments given in this paper show that with well selected and weighted language features, MaxEnt models with exponential priors are significantly more suitable for the text subjectivity analysis task.

The occlusal plane in the facial context:inter-operator repeatability of a new three-dimensional method

The repeatability of a non-invasive digital protocol proposed to evaluate the three-dimensional（3D） position of the occlusal plane in the face is assessed.Dental virtual models and soft tissue facial morphology of 20 adult subjects were digitally integrated using a 3D stereophotogrammetric imaging system.The digital 3D coordinates of facial and dental landmarks were obtained by two different operators.Camper＇s（facial） and occlusal（dental） planes were individuated,and their 3D relationships were measured.The repeatability of the protocol was investigated and showed no significant differences in repeated digitizations.The angle between occlusal and Camper＇s planes was smaller than 26 in the frontal and horizontal projections.In the sagittal projection,the angle was observed to be,on average,4.9 6.The determined occlusal plane pitch,roll and yaw values show good agreement with previously published data obtained by different protocols.The current non-invasive method was repeatable,without inter-operator differences and can facilitate assessment of healthy subjects.

Suitability of VVC and HEVC for Video Telehealth Systems

Video compression in medical video streaming is one of the key technologies associated with mobile healthcare.Seamless delivery of medical video streams over a resource constrained network emphasizes the need of a video codec that requires minimum bitrates and maintains high perceptual quality.This paper presents a comparative study between High Efciency Video Coding(HEVC)and its potential successor Versatile Video Coding(VVC)in the context of healthcare.A large-scale subjective experiment comprising of twenty-four non-expert participants is presented for eight different test conditions in Full High Denition(FHD)videos.The presented analysis highlights the impact of compression artefacts on the perceptual quality of HEVC and VVC processed videos.Our results and ndings show that VVC clearly outperforms HEVC in terms of achieving higher compression,while maintaining high quality in FHD videos.VVC requires upto 40%less bitrate for encoding an FHD video at excellent perceptual quality.We have provided rate-quality curves for both encoders and a degree of overlap across both codecs in terms of perceptual quality.Overall,there is a 71%degree of overlap in terms of quality between VVC and HEVC compressed videos for eight different test conditions.