Yazar "Nur Yilmaz, Gokce" seçeneğine göre listele

Listeleniyor 1 - 6 / 6
  • Küçük Resim
    Öğe
    An Abstraction and Structural Information Based Depth Perception Evaluation Metric
    (Ieee, 2017) Nur Yilmaz, Gokce; Bayrak, Huseyin
    Developing reliable and efficient 3 Dimensional (3D) video depth perception evaluation metrics is currently a trending research topic for supporting the advancement of the 3D video services. This support can be proliferated by utilizing effective 3D video features while modeling these metrics. In this study, a Reduced Reference (RR) depth perception evaluation metric using significant depth level and structural information as effective 3D video features is developed. The significant depth level and structural information in the Depth Maps (DM) are determined using abstraction filter and Canny edge detection algorithm, respectively. The performance assessment results of the proposed RR metric present that it is quite effective for ensuring advanced 3D video services.
  • Küçük Resim
    Öğe
    A Depth Perception Evaluation Metric For Immersive 3d Video Services
    (Ieee, 2017) Nur Yilmaz, Gokce
    Burgeoning advances in 3 Dimensional (3D) video services provide a big leap on the proliferation of the investigations for developing reliable and competent perceptual 3D video Quality of Experience (QoE) metrics. This proliferation can only be supported by exploiting key features characterizing 3D video nature in these investigations. In this paper, a Reduced Reference (RR) metric is developed considering that the spatial resolution and perceptually significant depth level are two effective features for efficiently evaluating depth perception of the 3D video. In order to determine the perceptually significant depth levels in the depth map sequences, abstraction filter is exploited in the development of the proposed metric. Owing to the fact that the depth perception significantly differs for the depth map sequences having dissimilar relative depth levels, this feature is also incorporated with the proposed metric through normalized standard deviation. Structural SIMilarity metric (SSIM) is utilized to predict the depth perception degraded with the change in the perceptually important levels of the compressed depth maps having dissimilar spatial resolutions and relative depth levels. The performance assessment of the proposed RR metric proves the effectiveness of the proposed metric for ensuring immersive 3D video services.
  • Küçük Resim
    Öğe
    Depth Perception Prediction of 3D Video for Ensuring Advanced Multimedia Services
    (Ieee, 2018) Nur Yilmaz, Gokce; Battisti, Federica
    A key role in the advancement of 3 Dimensional TV services is played by the development of 3D video quality metrics used for the assessment of the perceived quality. Moreover, this key role can only be supported when the features associated with the 3D video nature is reliably and efficiently characterized in these metrics. In this study, z-direction motion incorporated with significant depth levels in depth map sequences are considered as the main characterizations of the 3D nature. The 3D video quality metrics can be classified into three categories based on the need for the reference video during the assessment process at the user end: Full Reference (FR), Reduced Reference (RR) and No Reference (NR). In this study we propose a NR quality metric, PNRM, suitable for on-the-fly 3D video services. In order to evaluate the reliability and effectiveness of the proposed metric, subjective experiments are conducted in this paper. Observing the high correlation with the subjective experimental results, it can be clearly stated that the proposed metric is able to mimic the Human Visual System (HVS).
  • Küçük Resim
    Öğe
    Depth Perception Prediction of 3D Video QoE for Future Internet Services
    (Ieee, 2018) Nur Yilmaz, Gokce
    3 Dimensional (3D) video Quality of Experience (QoE) metrics are at utmost importance to enable enhancement of Future Internet Services. This enhancement can only be supported when the 3D video is characterized in the most reliable way as possible in these metrics. In light of this fact, a QoE metric including significant depth level and aerial perspective cue to support this reliable way is developed to predict the depth perception of the 3D video. Considering that No Reference (NR) QoE metric type is the most efficient one compared to the other types (i.e., Full Reference (FR) and Reduced Reference (RR)) in terms of transmission requirement, it is used as the metric type to develop the proposed metric. Conducted subjective experiments which are currently the "gold standard" in terms of reliable depth perception assessment are exploited to assess the performance of the proposed metric. Observing the effectiveness of the performance results, it can be clearly concluded that the advancement of the 3D video communication technologies can be ensured to assist Future Internet Services in a timely fashion.
  • Küçük Resim
    Öğe
    No-Reference Evaluation of 3 Dimensional Video Quality Using Spatial and Frequency Domain Components
    (Ieee, 2018) Bayrak, Huseyin; Nur Yilmaz, Gokce
    Video Quality Assessment (VQA) plays an important role both for evaluating the performance of the transmitter-receiver system and for delivering the video in an efficient manner via the feedback it provides to the transmitter side. Full Reference (FR) VQA metrics currently utilized in the literature are not too efficient during the applications due to the requirement of the original video sequence at the receiver side. Therefore, the tendency of the researchers is recently on to develop Reduced Reference (RR) or No-Reference (NR) VKD metrics. In this paper, a NR VKD metric considering spatial and frequency domain components of the color and depth map based 3 Dimensional (3D) video important for Human Visual System (HVS) is developed. Canny operator which is an efficient algorithm to extract edge information is used to obtain the components in the spatial domain. Discrete Cosine Transform (DCT) is exploited to obtain the components in the frequency domain. The efficient results obtained show that the proposed algorithm is capable of superseding the FR metrics existing in the literature.
  • Küçük Resim
    Öğe
    Scene Detection via Depth Maps Of 3 Dimensional Videos
    (Ieee, 2017) Bayrak, Huseyin; Nur Yilmaz, Gokce
    Scene detection via processing of multimedia data is a significant research area for the advancement of the video technologies and applications. Currently, the scene detection is mostly performed manually. Thus, it is time consuming and costly. Therefore, it is important to develop algorithms that can automatically segment scenes to support the advancement of these technologies and applications. With the wide-spread utilization of the 3 Dimensional (3D) videos, researchers working in the field of the video scene detection start using them in this field as well. However, there is still a gap in the application of the scene detection algorithms to Depth Maps (DMs) that are a part of the 3D video and important for temporal video scene detection. In this study, dominant clusters and K-means method is proposed to detect the temporal 3D video segments using the DMs. The experimental studies performed using the scene detection method present that the video scenes can be edited efficiently without human assistance. Moreover, unlike similar studies in the literature, the proposed method can provide successful results on video sequences thanks to the dominant clusters and the K-means clustering approach utilized.