Guiwon Seo

Auto-stereoscopic 3D displays with reduced crosstalk

In this paper, we propose new auto-stereoscopic 3D displays that substantially reduce crosstalk. In general, it is difficult to eliminate crosstalk in auto-stereoscopic 3D displays. Ideally, the parallax barrier can eliminate crosstalk for a single viewer at the ideal position. However, due to variations in the viewing distance and the interpupillary distance, crosstalk is a problem in parallax barrier displays. In this paper, we propose 3-dimensional barriers, which can significantly reduce crosstalk.

Video quality measurement for multimedia applications using reduced-reference signals

As more multimedia services have become increasingly available over networks where bandwidth is not always guaranteed, quality monitoring has become an important issue. For instance, quality of experience and quality monitoring have become important problems in internet protocol television applications, since transmission errors may introduce all kinds of additional video quality degradations. In this paper, we present a reduced-reference objective model for video quality measurements in multimedia applications. The proposed method first measures edge degradations that are critical for perceptual video quality and then considers transmission error effects. We compared the proposed method with some existing methods. Independent verifications confirmed that the proposed method showed good performance and consequently it was included in an International Telecommunication Union recommendation. The proposed method can be used to monitor video quality at receivers while requiring minimum usage of additional bandwidth.

Efficient implementation of neural network deinterlacing

Interlaced scanning has been widely used in most broadcasting systems. However, there are some undesirable artifacts such as jagged patterns, flickering, and line twitters. Moreover, most recent TV monitors utilize flat panel display technologies such as LCD or PDP monitors and these monitors require progressive formats. Consequently, the conversion of interlaced video into progressive video is required in many applications and a number of deinterlacing methods have been proposed. Recently deinterlacing methods based on neural network have been proposed with good results. On the other hand, with high resolution video contents such as HDTV, the amount of video data to be processed is very large. As a result, the processing time and hardware complexity become an important issue. In this paper, we propose an efficient implementation of neural network deinterlacing using polynomial approximation of the sigmoid function. Experimental results show that these approximations provide equivalent performance with a considerable reduction of complexity. This implementation of neural network deinterlacing can be efficiently incorporated in HW implementation.

Paper check image quality enhancement with Moire reduction

Paper checks may have complex background features including fine lines and patterns, which make forgery more difficult. Also, halftoning techniques are used to produce continuous tones and to prevent copies with void pantograph features. When these kinds of checks are scanned, Moire patterns may occur. These patterns make it difficult for customers to examine the scanned check images on ATM (Automated Teller Machine) displays. They also can decrease the classification accuracy of check recognition systems. In this paper, we propose an algorithm to enhance the perceptual quality of scanned check images by reducing the Moire patterns. The proposed algorithm consists of foreground extraction, Moire detection and Moire removal. Subjective image quality assessment was performed to evaluate the degree of improvement. Experimental results show that the proposed algorithm improves perceptual quality while maintaining check recognition accuracy.

Objective perceptual quality assesment and depth perception of 3D media

In this paper, we investigate the applicability of existing objective perceptual video quality measurement methods, which were developed for 2D video, to measure the perceptual quality of 3D video programs. We tested two methods: a reduced-reference HDTV model (ITU-T Rec. J.342) and a hybrid NR method (ITU-T Rec. J.343). As more mobile devices support autostereoscopic displays, crosstalk emerges as an important issue. We also investigate the effects of crosstalk on depth perception. Experimental results show promising performance and it appears that 2D video quality measurement methods can be used for 3D contents with potential improvements by taking into account 3D characteristics.

Objective video quality measurement using embedded VQMs

Video quality monitoring has become an important issue as multimedia data is increasingly being transmitted over the Internet and wireless channels where transmission errors can frequently occur. Although no-reference models are suitable to such applications, current no-reference methods do not provide acceptable performance. In this paper, we propose an objective video quality assessment method using embedded video quality metrics (VQMs). In the proposed method, the video quality of encoded video data is computed at the transmitter during the encoding process. The computed VQMs are embedded in the compressed data. If there are no transmission errors, the video quality at the receiver would be identical with that of the transmitting side. If there are transmission errors, the receiver adjusts the embedded VQMs by taking into account the effects of transmission errors. The proposed method is fast and provides good performance.

Frequency sensitivity for video compression

Abstract. We investigate the frequency sensitivity of the human visual system, which reacts differently at different frequencies in video coding. Based on this observation, we used different quantization steps for different frequency components in order to explore the possibility of improving coding efficiency while maintaining perceptual video quality. In other words, small quantization steps were used for sensitive frequency components while large quantization steps were used for less sensitive frequency components. We performed subjective testing to examine the perceptual video quality of video sequences encoded by the proposed method. The experimental results showed that a reduction in bitrate is possible without causing a decrease in perceptual video quality.

Direct Encoding for Sampled Color Pictures with Location Consideration

Single image sensors with a color filter array (CFA) have been widely used in digital cameras, digital camcorders, and smart phones. These sensors generate sampled RGB color images and interpolation produces full resolution RGB color images. Typically, color format conversion is applied to convert the RGB images to Yuv images and these Yuv images are then encoded. However, the interpolation process followed by color format conversion increases the data size by 50~100%. In order to avoid this potential inefficiency, it was proposed to encode sampled RGB color images without interpolation. However, due to location mismatching, coding improvement has not been consistent. In this paper, we consider the location differences in intra-mode prediction and preliminary experiments show that consistent improvement was obtained.

Fast implementation of neural network classification

Most artificial neural networks use a nonlinear activation function that includes sigmoid and hyperbolic tangent functions. Most artificial networks employ nonlinear functions such as these sigmoid and hyperbolic tangent functions, which incur high complexity costs, particularly during hardware implementation. In this paper, we propose new polynomial approximation methods for nonlinear activation functions that can substantially reduce complexity without sacrificing performance. The proposed approximation methods were applied to pattern classification problems. Experimental results show that the processing time was reduced by up to 50% without any performance degradations in terms of computer simulation.

Multimedia No-reference Video Quality Assessment Methods Using Bit Stream Information

Various video services with networks are increasingly available as smart phones, computers and IPTV are widely used. However, transmission over networks may experience transmission errors due to traffic increases and noise. As a result, video quality may suffer. Therefore, quality monitoring emerges as an important issue. In this paper, we propose a video quality assessment method using bit stream information. The video quality metric is calculated using header information and ES (elementary stream) information. To assess performance of the proposed algorithm, subjective quality assessment tests are conducted (VGA resolution). It is shown high correlation between subjective result and the proposed method.