# No-Reference Objective Video Quality Measure for Frame Freezing Degradation

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

- full-reference (FR) quality measures;
- reduced-reference (RR) quality measures;
- no-reference (NR) quality measures.

## 2. Related Work

- NR-FFM measure based on overall frame duration, number of frame freezings, and spatial information of the tested video sequence;
- NR-FFM obtains best correlation among other tested frame freezing objective measures in LIVE mobile dataset and similar correlation in VQEG HD5 dataset as other tested measures;
- NR-FFM can be combined with other objective measures designed for other degradation types, tested on LIVE mobile dataset with five degradation types, and combined with RVQM and STRRED as reduced reference measures.

## 3. Description of Used Datasets with Subjective Ratings

#### 3.1. LIVE Mobile Dataset

- H.264 compression (four test videos per reference);
- Wireless channel packet-loss (four test videos per reference);
- Frame-freezes (four test videos per reference);
- Rate adaptation (three test videos per reference);
- Temporal dynamics (five test videos per reference).

- DMOS score was lower (better) for longer frame freezing in the case of the stored streaming scenario (degradation types 1–3) with the same affected frame rate;
- DMOS score was higher (worse) in the case of live streaming scenario (1 × 120 + 1 × 60 frame freeze) where frames were ‘lost’, compared with the case of stored video and 2 × 120 frame freezes; however, it was not clear by how much, as the former had a freeze at the end lasting for 60 frames;
- DMOS score was different for different video sequences and the same degree of AFR, which means that the final DMOS score depended also upon video content—a good example would be sequence number 11, with a very low DMOS score, as this sequence had low spatial and temporal information, so the long frame freeze did not have a higher impact on the subjective score.

#### 3.2. VQEG HD5 Dataset

- hrc10: H.264/AVC compression, bitrate 16 Mbit/s, 1-pass encoding, bursty packet loss, packet loss rate 0.125%; freezing errors;
- hrc11: H.264/AVC compression, bitrate 16 Mbit/s, 1-pass encoding, bursty packet loss, packet loss rate 0.25%; freezing errors;
- hrc12: H.264/AVC compression, bitrate 16 Mbit/s, 1-pass encoding, bursty packet loss, packet loss rate 0.5%; freezing errors;
- hrc15: H.264/AVC compression, bitrate 4 Mbit/s, 1-pass encoding, bursty packet loss, packet loss rate 0.25%; freezing errors.

## 4. NR-FFM Measure Development

_{n}represents luminance plane at time n, Sobel represents the Sobel operator and by convolvingwith 3 × 3 kernel, calculation of the SI is defined as maximal value of all Sobel-filtered frames standard deviation (stdspace ) values. By default, Sobel operator should be calculated for both horizontal and vertical edges. However, we later calculated SI for only horizontal (later described as SI

_{H}), only vertical (later described as SI

_{V}), and both horizontal and vertical edges (later described as SI

_{H,V}):

## 5. Experiments and Results

#### 5.1. Results Using Frame Freezing Degradations—Overall Measure, LIVE Mobile Dataset

_{H}), both horizontal and vertical edges (SI

_{H,V}), and only vertical edges (SI

_{V}), respectively. Mean parameters α and β over non-overlapping test sets for those cases are presented in Table 4. After nonlinear regression using four parameter logistic functions Q

_{1}, Q

_{2,}Q

_{3}, and Q

_{4}, Pearson’s correlation was calculated according to:

_{1}and Q

_{2}are defined in [34] and [35], respectively, whereas Q

_{3}and Q

_{4}represent cubic and linear fit.

_{1}and Q

_{3}fitting functions. Q

_{2}had a somewhat lower correlation and Q

_{4}function gave the lowest correlation (as could be expected from linear fitting function).

_{1}, Q

_{2}, Q

_{3}, and Q

_{4}.

#### 5.2. NR-FFM Measure: Comparison between “Mobile” and “Tablet” Sub-Dataset from LIVE Mobile Dataset

_{1}, Q

_{2,}Q

_{3}, and Q

_{4}, respectively. Figure 4 shows the estimated values of NR-FFM measure versus the subjective DMOS scores using Equation (7). Also, Table 7 shows the fitting coefficients b

_{1}–b

_{5}, as defined in Equation (6), for this case.

_{1}, Q

_{2}, Q

_{3}, and Q

_{4}as fitting functions), also for this case.

#### 5.3. Combined Results, Overall LIVE Mobile Dataset

_{1}, Q

_{2}, and Q

_{3}have similar correlations, whereas Q

_{4}produces a much lower correlation. This means that the linear fitting function Q

_{4}cannot be used in this case. It has to be also noted that in this database, only one part from Equation (10) has an influence on the final objective metric (frame-freezing degradation or any other degradation type), whereas the other part shows perfect quality (with score 1).

#### 5.4. NR-FFM Measure Comparison with Other Objective Measures for Frame Freezing Degradations

_{H}(from Equations (1) and (2)) as spatial information values.

## 6. Discussion

## 7. Conclusions

## Author Contributions

## Funding

## Conflicts of Interest

## References

- ITU-R. BT.500-13 Methodology for the Subjective Assessment of the Quality of Television Pictures; International Telecommunication Union/ITU Radiocommunication Sector: Geneva, Switzerland, 2012. [Google Scholar]
- Video Quality Experts Group. Final Report from the Video Quality Experts Group on the Validation of Objective Models of Multimedia Quality. 2008. Available online: http://www.vqeg.org/ (accessed on 14 October 2019).
- Bjelopera, A.; Dumic, E.; Grgic, S. Evaluation of Blur and Gaussian Noise Degradation in Images Using Statistical Model of Natural Scene and Perceptual Image Quality Measure. Radioengineering
**2017**, 26, 930–937. [Google Scholar] [CrossRef] - Chikkerur, S.; Sundaram, V.; Reisslein, M.; Karam, L.J. Objective video quality assessment methods: A classification, review, and performance comparison. IEEE Trans. Broadcast.
**2011**, 57, 165–182. [Google Scholar] [CrossRef] - Loncaric, M.; Tralic, D.; Brzica, M.; Vukovic, J.; Lovrinic, J.; Dumic, E.; Grgic, S. Testing picture quality in HDTV systems. In Proceedings of the 50th International Symposium ELMAR, Zadar, Croatia, 10–12 September 2008; pp. 5–8. [Google Scholar]
- Wang, Z.; Bovik, A.C.; Sheikh, H.R.; Simoncelli, E.P. Image Quality Assessment: From Error Visibility to Structural Similarity. IEEE Trans. Image Process.
**2004**, 13, 600–612. [Google Scholar] [CrossRef] [PubMed][Green Version] - Chandler, D.M.; Hemami, S.S. VSNR: A Wavelet-Based Visual Signal-to-Noise Ratio for Natural Images. IEEE Trans. Image Process.
**2007**, 16, 2284–2298. [Google Scholar] [CrossRef] [PubMed] - Haque, M.I.; Qadri, M.T.; Siddiqui, N. Reduced Reference Blockiness and Blurriness Meter for Image Quality Assessment. Imaging Sci. J.
**2015**, 63, 296–302. [Google Scholar] [CrossRef] - Seshadrinathan, K.; Bovik, A.C. Motion Tuned Spatio-temporal Quality Assessment of Natural Videos. IEEE Trans. Image Process.
**2010**, 19, 335–350. [Google Scholar] [CrossRef] [PubMed] - Pinson, M.H.; Wolf, S. A new standardized method for objectively measuring video quality. IEEE Trans. Broadcast.
**2004**, 50, 312–322. [Google Scholar] [CrossRef] - Soundararajan, R.; Bovik, A.C. Video Quality Assessment by Reduced Reference Spatio-temporal Entropic Differencing. IEEE Trans. Image Process.
**2013**, 23, 684–694. [Google Scholar] - Dumic, E.; Grgic, S. Reduced Video Quality Measure Based on 3D Steerable Wavelet Transform and Modified Structural Similarity Index. In Proceedings of the 55th International Symposium ELMAR, Zadar, Croatia, 25–27 September 2013; pp. 65–69. [Google Scholar]
- Wolf, S.M.; Pinson, H. Video Quality Model for Variable Frame Delay (VQM_VFD); The National Telecommunications and Information Administration: Boulder, CO, USA, 2011. [Google Scholar]
- Qi, Y.; Dai, M. The Effect of Frame Freezing and Frame Skipping on Video Quality. In Proceedings of the International Conference on Intelligent Information Hiding and Multimedia, Pasadena, CA, USA, 18–20 December 2006; pp. 423–426. [Google Scholar]
- Huynh-Thu, Q.; Ghanbari, M. No-reference temporal quality metric for video impaired by frame freezing artefacts. In Proceedings of the 16th IEEE International Conference on Image Processing (ICIP), Cairo, Egypt, 7–11 November 2009; pp. 2221–2224. [Google Scholar]
- You, J.; Hannuksela, M.M.; Gabbouj, M. An objective video quality metric based on spatiotemporal distortion. In Proceedings of the 16th IEEE International Conference on Image Processing (ICIP), Cairo, Egypt, 7–11 November 2009; pp. 2229–2232. [Google Scholar]
- Pinson, M.H.; Choi, L.K.; Bovik, A.C. Temporal Video Quality Model Accounting for Variable Frame Delay Distortions. IEEE Trans. Broadcast.
**2014**, 60, 637–649. [Google Scholar] [CrossRef] - Moorthy, A.K.; Choi, L.K.; Bovik, A.C.; de Veciana, G. Video Quality Assessment on Mobile Devices: Subjective, Behavioral and Objective Studies. IEEE J. Sel. Top. Signal Process.
**2012**, 6, 652–671. [Google Scholar] [CrossRef][Green Version] - Borer, S. A model of jerkiness for temporal impairments in video transmission. In Proceedings of the 2010 Second International Workshop on Quality of Multimedia Experience (QoMEX), Trondheim, Norway, 21–23 June 2010; pp. 218–223. [Google Scholar]
- Wolf, S. A no reference (NR) and reduced reference (RR) metric for detecting dropped video frames. In Proceedings of the Fourth International Workshop on Video Processing and Quality Metrics for Consumer Electronics (VPQM), Scottsdale, AZ, USA, 15–16 January 2009; pp. 1–6. [Google Scholar]
- Usman, M.A.; Usman, M.R.; Shin, S.Y. A no reference method for detection of dropped video frames in live video streaming. In Proceedings of the Eighth International Conference on Ubiquitous and Future Networks (ICUFN), Vienna, Austria, 5–8 July 2016; pp. 839–844. [Google Scholar]
- Usman, M.A.; Shin, S.Y.; Shahid, M.; Lövström, B. A no reference video quality metric based on jerkiness estimation focusing on multiple frame freezing in video streaming. IETE Tech. Rev.
**2017**, 34, 309–320. [Google Scholar] [CrossRef] - Usman, M.A.; Usman, M.R.; Shin, S.Y. A novel no-reference metric for estimating the impact of frame freezing artifacts on perceptual quality of streamed videos. IEEE Trans. Multimed.
**2018**, 20, 2344–2359. [Google Scholar] [CrossRef] - Usman, M.A.; Usman, M.R.; Shin, S.Y. The impact of temporal impairment on quality of experience (QOE) in Video streaming: A no reference (NR) subjective and objective study. Int. J. Comput. Electr. Autom. Control Inf. Eng.
**2015**, 9, 1570–1577. [Google Scholar] - Xue, Y.; Erkin, B.; Wang, Y. A novel no-reference video quality metric for evaluating temporal jerkiness due to frame freezing. IEEE Trans. Multimed.
**2014**, 17, 134–139. [Google Scholar] [CrossRef] - Babić, D.; Stefanović, D.; Vranješ, M.; Herceg, M. Real-time no-reference histogram-based freezing artifact detection algorithm for UHD videos. Multimed. Tools Appl.
**2019**, 1–23. [Google Scholar] [CrossRef] - Grbić, R.; Stefanović, D.; Vranješ, M.; Herceg, M. Real-time video freezing detection for 4K UHD videos. J. Real-Time Image Process.
**2019**, 1–15. [Google Scholar] - VQEG Final Report of HDTV Validation Test, 2010 VQEG. Available online: https://www.its.bldrdoc.gov/vqeg/projects/hdtv/hdtv.aspx (accessed on 14 October 2019).
- The Consumer Digital Video Library. Available online: www.cdvl.org (accessed on 14 October 2019).
- ITU-T. P.910: Subjective Video Quality Assessment Methods for Multimedia Applications; International Telecommunication Union: Geneva, Switzerland, 1999. [Google Scholar]
- Miettinen, K.; Neittaanmäki, P.; Mäkelä, M.M.; Périaux, J. Evolutionary Algorithms in Engineering and Computer Science; John Wiley & Sons Ltd.: Hoboken, NJ, USA, 1999. [Google Scholar]
- Hauke, J.; Kossowski, T. Comparison of Values of Pearson’s and Spearman’s Correlation Coefficient on the Same Sets of Data. In Proceedings of the MAT TRIAD 2007 Conference, Birmingham, UK, 13–15 September 2007; pp. 87–93. [Google Scholar]
- David Garson, G. Correlation (Statistical Associates “Blue Book” Series Book 3); Statistical Associates Publishing: Asheboro, NC, USA, 2013. [Google Scholar]
- Sheikh, H.R. Image Quality Assessment Using Natural Scene Statistics. Ph.D. Thesis, University of Texas at Austin, Austin, TX, USA, 2004. [Google Scholar]
- Larson, E.C.; Chandler, D.M. Most apparent distortion: Full-reference image quality assessment and the role of strategy. J. Electron. Imaging
**2010**, 19. [Google Scholar] - Chenouard, N.; Unser, M. 3D Steerable Wavelets in practice. IEEE Trans. Image Process.
**2012**, 21, 4522–4533. [Google Scholar] [CrossRef] [PubMed] - Soundararajan, R.; Bovik, A.C. RRED indices: Reduced reference entropic differencing for image quality assessment. IEEE Trans. Image Process.
**2012**, 21, 517–526. [Google Scholar] [CrossRef] [PubMed] - Wolf, S. A No Reference (NR) and Reduced Reference (RR) Metric for Detecting Dropped Video Frames, NTIA Technical Memorandum TM-09-456, October 2008. Available online: https://www.its.bldrdoc.gov/publications/details.aspx?pub=2493 (accessed on 14 October 2019).
- CVQM Source Code for Matlab. Available online: https://www.its.bldrdoc.gov/resources/video-quality-research/guides-and-tutorials/description-of-vqm-tools.aspx (accessed on 14 October 2019).

**Figure 1.**First frame, laboratory for image and video engineering (LIVE) mobile database: (

**a**) bulldozer with fence, (

**b**) Barton Springs pool diving, (

**c**) friend drinking Coke, (

**d**) Harmonicat, (

**e**) landing airplane, (

**f**) panning under oak, (

**g**) Runners skinny guy, (

**h**) two swans dunking, (

**i**) students looming across street, (

**j**) trail pink kid.

**Figure 3.**Temporal information versus spatial information (SI

_{H,V}) for: (

**a**) LIVE mobile dataset, stored transmission scenario (blue) and live transmission scenario (red); (

**b**) VQEG dataset (video sequences src01, src02, src04, src05, src06, src08, and src09 with degradations hrc10, hrc11, hrc12, and hrc15, previously described).

**Figure 4.**No-reference frame–freezing measure (NR-FFM) versus DMOS (trained using LIVE mobile sub-dataset): (

**a**) LIVE mobile dataset; (

**b**) LIVE tablet dataset.

**Figure 5.**Combined objective measure: (

**a**) RVQM (reduced video quality measure), mobile dataset; (

**b**) STRRED (spatio-temporal reduced reference entropic differences), mobile dataset; (

**c**) RVQM, tablet dataset; (

**d**) STRRED, tablet dataset.

**Table 1.**Kendall’s, Spearman’s, and Pearson’s correlation for different train-test dataset ratio—SI

_{H}.

Train-Test Ratio | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation, Q_{1} | Pearson’s Correlation, Q_{2} | Pearson’s Correlation, Q_{3} | Pearson’s Correlation, Q_{4} | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|

Mean | Median | Mean | Median | Mean | Median | Mean | Median | Mean | Median | Mean | Median | |

7–3 | 0.6309 | 0.6364 | 0.7878 | 0.8042 | 0.8761 | 0.8809 | 0.8542 | 0.8619 | 0.8773 | 0.8838 | 0.8294 | 0.8381 |

8–2 | 0.6592 | 0.6429 | 0.7960 | 0.8095 | 0.9090 | 0.9018 | 0.8903 | 0.8820 | 0.9110 | 0.9133 | 0.8454 | 0.8651 |

**Table 2.**Kendall’s, Spearman’s, and Pearson’s correlation for different train-test dataset ratio—SI

_{H,V}.

Train-Test Ratio | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation, Q_{1} | Pearson’s Correlation, Q_{2} | Pearson’s Correlation, Q_{3} | Pearson’s Correlation, Q_{4} | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|

Mean | Median | Mean | Median | Mean | Median | Mean | Median | Mean | Median | Mean | Median | |

7–3 | 0.6433 | 0.6364 | 0.7964 | 0.7980 | 0.8768 | 0.8808 | 0.8620 | 0.8603 | 0.8811 | 0.8834 | 0.8328 | 0.8422 |

8–2 | 0.6796 | 0.6429 | 0.8101 | 0.8095 | 0.9008 | 0.8983 | 0.8882 | 0.8894 | 0.9109 | 0.9203 | 0.8608 | 0.8680 |

**Table 3.**Kendall’s, Spearman’s, and Pearson’s correlation for different train-test dataset ratio—SI

_{V}.

Train-Test Ratio | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation, Q_{1} | Pearson’s Correlation, Q_{2} | Pearson’s Correlation, Q_{3} | Pearson’s Correlation, Q_{4} | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|

Mean | Median | Mean | Median | Mean | Median | Mean | Median | Mean | Median | Mean | Median | |

7–3 | 0.6536 | 0.6667 | 0.8089 | 0.8112 | 0.8887 | 0.8914 | 0.8705 | 0.8710 | 0.8882 | 0.8930 | 0.8396 | 0.8447 |

8–2 | 0.6869 | 0.6429 | 0.8144 | 0.8095 | 0.9095 | 0.9078 | 0.8927 | 0.8884 | 0.9098 | 0.9099 | 0.8550 | 0.8620 |

**Table 4.**Mean parameters α and β over non-overlapping test sets. SI

_{H}: only horizontal spatial information, SI

_{H,V}: horizontal and vertical spatial information, SI

_{V}: only vertical spatial information.

Train-Test Ratio | SI_{H} | SI_{H,V} | SI_{V} | |||
---|---|---|---|---|---|---|

Mean α | Mean β | Mean α | Mean β | Mean α | Mean β | |

7–3 | 0.5531 | 0.1769 | 0.4962 | 0.1211 | 0.4727 | 0.1308 |

8–2 | 0.5861 | 0.1550 | 0.5407 | 0.0962 | 0.4721 | 0.1446 |

**Table 5.**Kendall’s, Spearman’s, and Pearson’s correlation for the overall mobile dataset, different SI calculations (best values are marked in bold). H: horizontal, V: vertical.

SI Type | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation Using Q_{1} | Pearson’s Correlation Using Q_{2} | Pearson’s Correlation Using Q_{3} | Pearson’s Correlation Using Q_{4} | α | β |
---|---|---|---|---|---|---|---|---|

H only | 0.7026 | 0.8842 | 0.8952 | 0.8840 | 0.8978 | 0.8868 | 0.6327 | 0.1167 |

H and V | 0.6795 | 0.8598 | 0.8919 | 0.8790 | 0.8936 | 0.8574 | 0.5824 | 0.1672 |

V only | 0.6846 | 0.8683 | 0.8882 | 0.8677 | 0.8913 | 0.8353 | 0.2917 | 0.2127 |

**Table 6.**Kendall’s, Spearman’s, and Pearson’s correlation, using LIVE mobile, stored transmission video sequences, and LIVE mobile live transmission video sequences.

LIVE Dataset | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation Using Q_{1} | Pearson’s Correlation Using Q_{2} | Pearson’s Correlation Using Q_{3} | Pearson’s Correlation Using Q_{4} |
---|---|---|---|---|---|---|

Stored transmission | 0.7103 | 0.8919 | 0.9181 | 0.9084 | 0.9226 | 0.8906 |

Live transmission | 0.3333 | 0.3576 | 0.6526 | 0.6150 | 0.7008 | 0.4334 |

Fitted Function | Mobile Dataset | Tablet Dataset | ||||||||
---|---|---|---|---|---|---|---|---|---|---|

b_{1} | b_{2} | b_{3} | b_{4} | b_{5} | b_{1} | b_{2} | b_{3} | b_{4} | b_{5} | |

Q_{1} | 84 | 123.8 | 0.04294 | 15.1 | −41.17 | 18.72 | 16.42 | 0.1091 | −62.74 | 9.021 |

Q_{2} | −500.9 | 3.758 | −0.2237 | 0.05604 | - | 1.442 | 2.818 | 0.1054 | 0.02429 | - |

Q_{3} | 9429 | −3667 | 482.7 | −18.74 | - | −1687 | 555.6 | −46.82 | 2.863 | - |

Q_{4} | 18.07 | 0.4138 | - | - | - | 10.98 | 0.9761 | - | - | - |

**Table 8.**Kendall’s, Spearman’s, and Pearson’s correlation for the overall frame freezing dataset, using the LIVE tablet sub-dataset for training and the LIVE mobile sub-dataset for testing.

LIVE Dataset | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation Using Q_{1} | Pearson’s Correlation Using Q_{2} | Pearson’s Correlation Using Q_{3} | Pearson’s Correlation Using Q_{4} |
---|---|---|---|---|---|---|

Tablet | 0.6000 | 0.8030 | 0.8422 | 0.8417 | 0.8422 | 0.8312 |

Mobile | 0.7026 | 0.8842 | 0.8957 | 0.8838 | 0.8978 | 0.8668 |

**Table 9.**Kendall’s, Spearman’s, and Pearson’s correlation for different train-test dataset ratio (best values are marked in bold).

Measure | Train-Test Ratio | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation, Q_{1} | Pearson’s Correlation, Q_{2} | Pearson’s Correlation, Q_{3} | Pearson’s Correlation, Q_{4} | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|

Mean | Median | Mean | Median | Mean | Median | Mean | Median | Mean | Median | Mean | Median | ||

RVQM | 7–3 | 0.6255 | 0.6193 | 0.8065 | 0.7995 | 0.8390 | 0.8383 | 0.8290 | 0.8262 | 0.8272 | 0.8281 | 0.7019 | 0.7052 |

8–2 | 0.6494 | 0.6692 | 0.8210 | 0.8396 | 0.8669 | 0.8859 | 0.8527 | 0.8800 | 0.8494 | 0.8804 | 0.7012 | 0.7120 | |

STRRED | 7–3 | 0.7115 | 0.7173 | 0.8773 | 0.8852 | 0.9131 | 0.9166 | 0.9059 | 0.9122 | 0.8895 | 0.8918 | 0.7049 | 0.7083 |

8–2 | 0.7204 | 0.7283 | 0.8785 | 0.8861 | 0.9219 | 0.9256 | 0.9150 | 0.9214 | 0.8982 | 0.9044 | 0.7004 | 0.7043 |

**Table 10.**Kendall’s, Spearman’s, and Pearson’s correlation, before and after combining RVQM and STRRED with freezing degradations; VQM-VFD-RR and VQM-VFD-FR correlation from [17]; mobile dataset (best values are marked in bold).

Measure | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation Using Q_{1} | Pearson’s Correlation Using Q_{2} | Pearson’s Correlation Using Q_{3} | Pearson’s Correlation Using Q_{4} | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|

Before Merging (160 Videos) | After Merging (200 Videos) | Before Merging (160 Videos) | After Merging (200 Videos) | Before Merging (160 Videos) | After Merging (200 Videos) | Before Merging (160 Videos) | After Merging (200 Videos) | Before Merging (160 Videos) | After Merging (200 Videos) | Before Merging (160 Videos) | After Merging (200 Videos) | |

RVQM | 0.5779 | 0.6129 | 0.7697 | 0.8034 | 0.7790 | 0.8045 | 0.7780 | 0.8032 | 0.7780 | 0.8039 | 0.7108 | 0.7142 |

STRRED | 0.7214 | 0.7237 | 0.8902 | 0.8929 | 0.9125 | 0.9156 | 0.9062 | 0.9085 | 0.8921 | 0.8936 | 0.7542 | 0.7275 |

VQM-VFD-RR | - | - | - | 0.8301 | - | 0.8645 | - | - | - | - | - | - |

VQM-VFD-FR | - | - | - | 0.8295 | - | 0.8631 | - | - | - | - | - | - |

**Table 11.**Kendall’s, Spearman’s, and Pearson’s correlation, before and after combining RVQM and STRRED with freezing degradations; VQM-VFD-RR and VQM-VFD-FR correlation from [17]; tablet dataset (best values are marked in bold).

Measure | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation Using Q_{1} | Pearson’s Correlation Using Q_{2} | Pearson’s Correlation Using Q_{3} | Pearson’s Correlation Using Q_{4} | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|

Before Merging (80 Videos) | After Merging (100 Videos) | Before Merging (80 Videos) | After Merging (100 Videos) | Before Merging (80 Videos) | After Merging (100 Videos) | Before Merging (80 Videos) | After Merging (100 Videos) | Before Merging (80 Videos) | After Merging (100 Videos) | Before Merging (80 Videos) | After Merging (100 Videos) | |

RVQM | 0.5128 | 0.4972 | 0.6704 | 0.6621 | 0.7236 | 0.7362 | 0.7192 | 0.7261 | 0.7244 | 0.6991 | 0.6621 | 0.5206 |

STRRED | 0.6287 | 0.5853 | 0.8104 | 0.7731 | 0.8844 | 0.8645 | 0.8648 | 0.8563 | 0.8418 | 0.7848 | 0.6814 | 0.5433 |

VQM-VFD-RR | - | - | - | 0.8133 | - | 0.8110 | - | - | - | - | - | - |

VQM-VFD-FR | - | - | - | 0.8385 | - | 0.8347 | - | - | - | - | - | - |

**Table 12.**Kendall’s, Spearman’s, and Pearson’s correlation for the LIVE mobile dataset—40 degraded video sequences (best values are marked in bold). Borer: no reference objective measure, FDF: no reference objective measure, FDF-RR: reduced reference objective measure, Quanhuyn: no reference objective measure.

Measure | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation Using Q_{1} | Pearson’s Correlation Using Q_{2} | Pearson’s Correlation Using Q_{3} | Pearson’s Correlation Using Q_{4} |
---|---|---|---|---|---|---|

NR-FFM_8:2 | 0.6592 | 0.7960 | 0.9090 | 0.8903 | 0.9110 | 0.8454 |

Borer | 0.3663 | 0.5251 | 0.8286 | 0.7216 | 0.8232 | 0.0896 |

FDF | 0.0725 | 0.0873 | 0.2590 | 0.2553 | 0.2590 | 0.2064 |

FDF-RR | 0.0290 | 0.0375 | 0.2590 | 0.1572 | 0.2590 | 0.0790 |

Quanhuyn | 0.3203 | 0.4594 | 0.7443 | 0.6221 | 0.7443 | 0.4253 |

**Table 13.**Kendall’s, Spearman’s, and Pearson’s correlation for the VQEG dataset—28 degraded video sequences (best values are marked in bold).

Measure | Kendall’s Correlation | Spearman’s Correlation | Pearson’s Correlation Using Q_{1} | Pearson’s Correlation Using Q_{2} | Pearson’s Correlation Using Q_{3} | Pearson’s Correlation Using Q_{4} |
---|---|---|---|---|---|---|

NR-FFM | 0.6596 | 0.8523 | 0.8845 | 0.8736 | 0.8750 | 0.8425 |

Borer | 0.6916 | 0.8685 | 0.8736 | 0.8729 | 0.8735 | 0.8496 |

FDF | 0.6344 | 0.8240 | 0.8206 | 0.8187 | 0.8198 | 0.8072 |

FDF-RR | 0.6756 | 0.8576 | 0.8527 | 0.8534 | 0.8529 | 0.8372 |

Quanhuyn | 0.4147 | 0.5129 | 0.6056 | 0.5592 | 0.5900 | 0.4468 |

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Dumic, E.; Bjelopera, A.
No-Reference Objective Video Quality Measure for Frame Freezing Degradation. *Sensors* **2019**, *19*, 4655.
https://doi.org/10.3390/s19214655

**AMA Style**

Dumic E, Bjelopera A.
No-Reference Objective Video Quality Measure for Frame Freezing Degradation. *Sensors*. 2019; 19(21):4655.
https://doi.org/10.3390/s19214655

**Chicago/Turabian Style**

Dumic, Emil, and Anamaria Bjelopera.
2019. "No-Reference Objective Video Quality Measure for Frame Freezing Degradation" *Sensors* 19, no. 21: 4655.
https://doi.org/10.3390/s19214655