Learn More
Traditional isolated monolingual name taggers tend to yield inconsistent results across two languages. In this paper, we propose two novel approaches to jointly and consistently extract names from parallel corpora. The first approach uses standard linear-chain Conditional Random Fields (CRFs) as the learning framework, incorporating cross-lingual features(More)
Facial video sequences are encoded at very low bi-trates through the use of principal component analysis. All facial emotions can be generated by blending six basic emotions. Principal components are used to model the basic facial mimic and encode facial video frames into a few parameters, which can be transmitted over low capacity networks. Hands-free(More)
Universal Multimedia Access (UMA) engines are used to adapt the media content to suite different environments. The adaptation decision for content used over wireless networks is heavily dependent on the bandwidth needs since the transmission cost over wireless networks is high. We customize lip video into lip animations and through tests we challenge the(More)
A full-frame compression technique for video sequences of faces, based on principal component analysis, is presented. Entire video frames are encoded to very low bi-trates and reconstructed with good accuracy. Principal component analysis is utilized to model the facial mimic in an Eigenspace. By using a camera construction where the user's face is almost(More)
Word Segmentation is usually considered an essential step for many Chinese and Japanese Natural Language Processing tasks, such as name tagging. This paper presents several new observations and analysis on the impact of word segmentation on name tagging; (1). Due to the limitation of current state-of-the-art Chinese word segmentation performance, a(More)
In this paper, a pilot-based blind (PBB) phase estimation method for coherent optical orthogonal frequency-division multiplexing (CO-OFDM) is proposed. Instead of inserting the pilot-subcarriers that are loaded with the already known information in the OFDM signal, the unknown and specially designed signal is used to replace the signal on the pilot(More)
This master thesis in media technology depicts the development on the smartphone market globally but with focus on Swedish users. It underlines the importance of performance and processing power of smartphones on the Swedish market. In this study, user priorities when choosing a new smartphone are examined as well as user habits and user experience. The(More)
Using the shear strength reduction(SSR) finite element method(FEM), the safety factor (SF) of a homogeneous slope was calculated, the results showed that the differences were inevitable because of the different adopted instability criteria and personal judgments, even if the same criterion was used. According to analyzing the characteristics of energy(More)
We propose a novel guard-band-shared direct detection scheme to improve the receiver spectrum efficiency. A 100-Gb/s DDO-OFDM/OQAM signal transmission has been successfully demonstrated over 80-km SSMF using a single photodiode within 50-GHz optical grid.