Stereoscopic Video Quality Assessment Based on The Two-step-training Binocular Fusion Network

Ma, Shuai; Li, Sumei; Xue, Jiang; Ding, Yixiu; Yue, Guanghui

doi:10.1109/vcip47243.2019.8965700

Cited by 9 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As the visual signal is transmitted in the visual pathway, there are many subtle changes in the signal processing method especially in binocular information interaction method and attention mechanism [12][13]. The method in [7] attempted to construct a fusion channel based on the above knowledge but ignored the changes of mechanism in the visual cortex.…”

Section: Multi-stage Growing Attention-guided Fusion Channelmentioning

confidence: 99%

“…In Table I, the performance of the proposed method on two databases is compared with the other six classical methods including manual feature-based methods [2][3][4] and CNNbased methods [5][6][7]. The top 1 performance for each index is highlighted in boldface and unavailable results are marked as '-'.…”

Section: Overall Evaluationmentioning

confidence: 99%

“…Yang et al [6] designed a 3D convolutional neural network for NR-SVQA. Ma et al [7] used a two-step-training strategy to train the proposed network and spatio-temporal salient feature flows were taken as input to obtain the final perceptual quality. Some methods for stereoscopic panoramic video quality assessment appeared recently [8][9].…”

Section: Introductionmentioning

confidence: 99%

“…Stereoscopic video has spatial and temporal information simultaneously. Some methods [6,7] attempt to capture motion information contained in the adjacent video frames by spatiotemporal saliency feature flows and 3D convolution. They both have proved that letting the model learn temporal domain information is an effective way.…”

Section: Introductionmentioning

confidence: 99%

“…After that binocular fusion and binocular rivalry occur hierarchically in the higher visual cortex (V2, V3, V3A, V4, V5, and so on). The method in [7] added a fusion channel with several same weight blocks to the model to simulate the above process. However, actually, the trigger features in different visual cortexes are disparate and the visual attention mechanism is varied.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Multi-Scale Feature-Guided Stereoscopic Video Quality Assessment Based on 3d Convolutional Neural Network

Feng

Chang

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

With the huge development of stereoscopic video technology, the research of stereoscopic video quality assessment (SVQA) has become very important for promoting the development of stereoscopic video system. These years, many SVQA methods based on convolutional neural network (CNN) have emerged. In this paper, we proposed a multi-scale feature-guided 3D convolutional neural network for SVQA which not only use 3D convolution to capture spatio-temporal features but also aggregate multi-scale information by a new multi-scale unit. Besides, we employ a multi-stage growing attention mechanism in this network to learn more critical deep semantic information. The proposed method is tested on two public stereoscopic video quality datasets, and the result shows that this method correlates highly with human visual perception and outperforms stateof-the-art methods by a large margin.

show abstract

Section: Multi-stage Growing Attention-guided Fusion Channelmentioning

confidence: 99%