No-Reference Quality Assessment for Stitched Panoramic Images Using Convolutional Sparse Coding and Compound Feature Selection

Ling, Suiyi; Cheung, Gene; Callet, Patrick Le

doi:10.1109/icme.2018.8486545

Cited by 34 publications

(23 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using simple and pixel-level assessments of panoramic images, the proposed method provides an accurate estimation of perceptual quality, thereby exploiting the disturbance of pixels only in distortion-specific regions rather than traversing the entire panoramic image. In order to validate the effectiveness and generalization of the proposed DLNR-SIQA framework, we conducted a comparative analysis with existing deep learning-based FR-SIQA and NR-SIQA approaches [31,33,34]. The comparison was performed on two publicly available stitched images datasets: the SIQA [31,33] and the ISIQA (Indian Institute of Science Stitched Image QA) [34] dataset.…”

Section: Distorted Region Extraction and Quality Estimationmentioning

confidence: 99%

“…Recently, several NR-SIQA methods [ 33 , 34 , 35 , 36 , 37 ] have been proposed to automate the SIQA process. These methods estimate the perceptual quality of a given stitched image without using any stimulus information.…”

Section: Introductionmentioning

confidence: 99%

“…These methods estimate the perceptual quality of a given stitched image without using any stimulus information. For example, in [ 33 ], the authors introduced a convolutional sparse coding (CSC) technique to learn the pattern of stitching relevant distortion in a target image. They used different sets of convolution filters to localize the distortion region and, later, quantified the compound effect of these localized distortions using trained kernels.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

DLNR-SIQA: Deep Learning-Based No-Reference Stitched Image Quality Assessment

Ullah

Irfan

Han

et al. 2020

Sensors

View full text Add to dashboard Cite

Due to recent advancements in virtual reality (VR) and augmented reality (AR), the demand for high quality immersive contents is a primary concern for production companies and consumers. Similarly, the topical record-breaking performance of deep learning in various domains of artificial intelligence has extended the attention of researchers to contribute to different fields of computer vision. To ensure the quality of immersive media contents using these advanced deep learning technologies, several learning based Stitched Image Quality Assessment methods have been proposed with reasonable performances. However, these methods are unable to localize, segment, and extract the stitching errors in panoramic images. Further, these methods used computationally complex procedures for quality assessment of panoramic images. With these motivations, in this paper, we propose a novel three-fold Deep Learning based No-Reference Stitched Image Quality Assessment (DLNR-SIQA) approach to evaluate the quality of immersive contents. In the first fold, we fined-tuned the state-of-the-art Mask R-CNN (Regional Convolutional Neural Network) on manually annotated various stitching error-based cropped images from the two publicly available datasets. In the second fold, we segment and localize various stitching errors present in the immersive contents. Finally, based on the distorted regions present in the immersive contents, we measured the overall quality of the stitched images. Unlike existing methods that only measure the quality of the images using deep features, our proposed method can efficiently segment and localize stitching errors and estimate the image quality by investigating segmented regions. We also carried out extensive qualitative and quantitative comparison with full reference image quality assessment (FR-IQA) and no reference image quality assessment (NR-IQA) on two publicly available datasets, where the proposed system outperformed the existing state-of-the-art techniques.

show abstract

Section: Distorted Region Extraction and Quality Estimationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

DLNR-SIQA: Deep Learning-Based No-Reference Stitched Image Quality Assessment

Ullah

Irfan

Han

et al. 2020

Sensors

View full text Add to dashboard Cite

show abstract

“…As per as objective quality of stitched image is concerned, very few research attention is paid to blind image quality assessment of panoramic view. Recently, Suiyi Ling et al proposed no reference convolutional sparse coding using trained kernel capture distortion in local region of stitched image [20].…”

Section: Review Of Related Workmentioning

confidence: 99%

Blind Image Quality Evaluation of Stitched Image using Novel Hybrid Warping Technique

Gandhe¹,

Vaidya²

2019

IJACSA

View full text Add to dashboard Cite

Image stitching is collection of sequential images captured at fixed camera center having considerable amount of overlap and produces aesthetically pleasing seamless panoramic view. But, practically it is very difficult to obtain clean and pristine stitched panoramic image of particular scene as such images are apparently distorted. In this paper, a novel Hybrid Warping technique is used that combine two global warps and one local warp and helps to refine image alignment stage. Our proposed method optimizes Homography Screening to rectify problem of perspective distortion and Edge Strength Similarity approach to quantify structural irregularities. The Blind Image Quality Evaluation models such as Blind Image Quality Index (BIQI), Blind/Reference-less Image Spatial QUality Evaluator (BRISQUE) and BLind Image Integrity Notator using DCT Statistics (BLIINDS-II) are employed to measure objective quality of stitched image. The experimental results showed that blind image quality score of proposed method is significantly better than latest existing methods.

show abstract

“…This interactive property enables users to feel like being in a virtual world. It gives rise to various new challenges at the same time, such as video/image production [1], [2], [3], transmission [4], [5], compression [6], [7], [8], and quality assessments [9], [10], [11]. Those new challenges are dissimilar to the cases in 2D traditional media since users can actively select the content they would like to watch with HMDs, while they are only allowed to passively receive the given content in 2D traditional video.…”

Section: Introductionmentioning

confidence: 99%

A Multi-FoV Viewport-Based Visual Saliency Model Using Adaptive Weighting Losses for 360$^\circ$ Images

Chao

Zhang

Hamidouche

et al. 2021

IEEE Trans. Multimedia

View full text Add to dashboard Cite

• media allows observers to explore the scene in all directions. The consequence is that the human visual attention is guided by not only the perceived area in the viewport but also the overall content in 360 •. In this paper, we propose a method to estimate the 360 • saliency map which extracts salient features from the entire 360 • image in each viewport in three different Field of Views (FoVs). Our model is first pretrained with a largescale 2D image dataset to enable the interpretation of semantic contents, then fine-tuned with a relative small 360 • image dataset. A novel weighting loss function attached with stretch weighted maps is introduced to adaptively weight the losses of three evaluation metrics and attenuate the impact of stretched regions in equirectangular projection during training process. Experimental results demonstrate that our model achieves better performance with the integration of three FoVs and its diverse viewport images. Results also show that the adaptive weighting losses and stretch weighted maps effectively enhance the evaluation scores compared to the fixed weighting losses solutions. Comparing to other state of the art models, our method surpasses them on three different datasets and ranks the top using 5 performance evaluation metrics on the Salient360! benchmark set.

show abstract

No-Reference Quality Assessment for Stitched Panoramic Images Using Convolutional Sparse Coding and Compound Feature Selection

Cited by 34 publications

References 16 publications

DLNR-SIQA: Deep Learning-Based No-Reference Stitched Image Quality Assessment

DLNR-SIQA: Deep Learning-Based No-Reference Stitched Image Quality Assessment

Blind Image Quality Evaluation of Stitched Image using Novel Hybrid Warping Technique

A Multi-FoV Viewport-Based Visual Saliency Model Using Adaptive Weighting Losses for 360$^\circ$ Images

Contact Info

Product

Resources

About