Current Issue : April - June Volume : 2018 Issue Number : 2 Articles : 5 Articles
We propose amodular no-reference video quality predictionmodel for videos that are encoded with H.265/HEVC and VP9 codecs\nand viewed on mobile devices. The impairments which can affect video transmission are classified into two broad types depending\nupon which layer of the TCP/IP model they originated from. Impairments from the network layer are called the network QoS\nfactors, while those from the application layer are called the application/payload QoS factors. Initially we treat the network and\napplication QoS factors separately and find out the 1 : 1 relationship between the respective QoS factors and the corresponding\nperceived video quality or QoE. The mapping from the QoS to the QoE domain is based upon a decision variable that gives an\noptimal performance. Next, across each group we choose multiple QoS factors and find out the QoE for such multifactor impaired\nvideos by using an additive, multiplicative, and regressive approach. We refer to these as the integrated network and application\nQoE, respectively. At the end, we use a multiple regression approach to combine the network and application QoE for building\nthe final model.We also use an Artificial Neural Network approach for building the model and compare its performance with the\nregressive approach....
Extracting information about academic activity transactions from unstructured documents is a key problem in the analysis of\nacademic behaviors of researchers. The academic activities transaction includes five elements: person, activities, objects, attributes,\nand time phrases. The traditional method of information extraction is to extract shallow text features and then to recognize\nadvanced features from text with supervision. Since the information processing of different levels is completed in steps, the error\ngenerated from various steps will be accumulated and affect the accuracy of final results. However, because Deep Belief Network\n(DBN) model has the ability to automatically unsupervise learning of the advanced features from shallow text features, the model\nis employed to extract the academic activities transaction. In addition, we use character-based feature to describe the raw features\nof named entities of academic activity, so as to improve the accuracy of named entity recognition. In this paper, the accuracy of the\nacademic activities extraction is compared by using character-based feature vector and word-based feature vector to express the\ntext features, respectively, and with the traditional text information extraction based on Conditional Random Fields. The results\nshow that DBN model is more effective for the extraction of academic activities transaction information....
This paper presents a novel Gabor phase based illumination invariant extraction method aiming at eliminating the effect of varying\nillumination on face recognition. Firstly, It normalizes varying illumination on face images, which can reduce the effect of varying\nillumination to some extent. Secondly, a set of 2D real Gabor wavelet with different directions is used for image transformation, and\nmultiple Gabor coefficients are combined into one whole in considering spectrum and phase. Lastly, the illumination invariant is\nobtained by extracting the phase feature from the combined coefficients. Experimental results on the Yale B and the CMU PIE face\ndatabase show that our method obtained a significant improvement over other related methods for face recognition under large\nillumination variation condition....
Automatic colorization is generally classified into two groups: propagation-based methods and reference-based methods. In\nreference-based automatic colorization methods, color image(s) are used as reference(s) to reconstruct original color of a gray\ntarget image. The most important task here is to find the best matching pairs for all pixels between reference and target images in\norder to transfer color information from reference to target pixels. A lot of attractive local feature-based image matching methods\nhave already been developed for the last two decades. Unfortunately, as far as we know, there are no optimal matching methods\nfor automatic colorization because the requirements for pixel matching in automatic colorization are wholly different from those\nfor traditional image matching. To design an efficient matching algorithm for automatic colorization, clustering pixel with low\ncomputational cost and generating descriptive feature vector are the most important challenges to be solved. In this paper, we\npresent a novel method to address these two problems. In particular, our work concentrates on solving the second problem\n(designing a descriptive feature vector); namely, we will discuss how to learn a descriptive texture feature using scaled sparse\ntexture feature combining with a nonlinear transformation to construct an optimal feature descriptor. Our experimental results\nshow our proposed method outperforms the state-of-the-art methods in terms of robustness for color reconstruction for automatic\ncolorization applications....
Internet services that share vehicle black box videos need a way to obfuscate license plates in uploaded videos because of privacy\nissues.Thus, plate detection is one of the critical functions that such services rely on. Even though various types of detection methods\nare available, they are not suitable for black box videos because no assumption about size, number of plates, and lighting conditions\ncan bemade.We propose amethod to detect Korean vehicle plates fromblack box videos. It works in two stages: the first stage aims\nto locate a set of candidate plate regions and the second stage identifies only actual plates fromcandidates by using a support vector\nmachine classifier. The first stage consists of five sequential substeps. At first, it produces candidate regions by combining single\ncharacter areas and then eliminates candidate regions that fail to meet plate conditions through the remaining substeps. For the\nsecond stage, we propose a feature vector that captures the characteristics of plates in texture and color. For performance evaluation,\nwe compiled our dataset which contains 2,627 positive and negative images.The evaluation results show that the proposed method\nimproves accuracy and sensitivity by at least 5% and is 30 times faster compared with an existing method....
Loading....