The biggest challenges in automation of testing of complex consumer devices nowadays are:
- motion detection
- region comparison and
- optical character recognition (OCR)
being critical for the majority of tests, as they replace the need for manual recognition and verification.
Video analysis of BBT automated test solutions comprises of various algorithms for real time video artifact detection integrated in its devices for video analysis (PSNR, PBC, OCR) and products RT-AV400 and RT-IP030 (PQM - Picture Quality Metric).
- Peak Signal to Noise Ratio (PSNR)
- Picture Block Compare (PBC) metric
- Optical Character Recognition (OCR)
Video Algorithms
Peak Signal-to-Noise Ratio (PSNR)
PSNR is the most widely used objective video quality metric. It represents the ratio between the maximum possible power of signal and the power of corrupting noise that affects the fidelity of its representation.
However, PSNR values do not perfectly correlate with perceived visual quality due to non-linear behavior of the human visual system, which resulted in development of more complicated and precise metrics.
Picture Block Compare (PBC) metric
With PBC metric, it has been aimed at developing an efficient objective scheme for detection of digital picture degradations in television within a fully automated fault diagnosis system.
In addition to detecting digital picture degradations like blocking, partial or total picture loss, freeze of moving element, the method is highly sensitive to non-uniform distortion over the image, e.g. when only a small region in an image is corrupted with significant artifacts, while all other regions have high quality.
PBC scheme outstands PSNR and some other metrics such as SSIM (Structural Similarity Index Measurement) and VIF (Visual Information Fidelity) in terms of better error localization, as well as robustness to luminance and contrast change or negligible picture misalignment.
For more information on the metric, and comparative analysis against other objective image quality measures, see Quality assessment of digital television picture based on local feature matching paper.
Optical Character Recognition (OCR)
BBT test suite contains powerful Optical Character Recognition (OCR) for testing of Sub-titles, TTX/MHEG and On Screen Display (OSD) features. It supports various file formats, region of interest, and variety of transformations for better text extraction.
See examples of OCR performance.
OCR’s full capacity has been brought to bear in RT-AV400 with use cases in verification on the final production line, functional, large scale and regression testing.
Video Artifacts
Here are examples of some common video artifacts detected by BBT:
Brightness Errors
Errors related to brightness can come from two different reasons:
- Default brightness can vary in different DTV/Set-up box sets due to erroneous parameters in software -> errors in API software layer.
- Changing the brightness within the device under test (DUT) can produce images of wrong brightness values (differing significantly from the referent one) -> false hardware/software functionality.
BBT analyzes the difference between the grabbed picture from the reference DTV/set-up-box device. Based on the expected/allowed difference it is decided whether the DUT is of acceptable functionality or not, regarding the brightness.
Color Errors
Errors related to color can come from two different reasons:
- Default color can vary in different DTV/Set-up box sets due to erroneous parameters in software. -> errors in API software layer.
- Changing the color within the device under test (DUT) can produce images of wrong color values (differing significantly from the referent one) -> false hardware/software functionality.
BBT analyzes the difference between the grabbed picture from the reference DTV/set-up-box device. Based on the expected/allowed difference it is decided whether the DUT is of acceptable functionality or not, regarding the color.
Sharpness and Contrast Errors
Errors related to sharpness/contrast can come from two different reasons:
- Default sharpness/contrast can vary in different DTV/Set-up box sets due to erroneous parameters in software. -> errors in API software layer.
- Changing the sharpness/contrast within the device under test (DUT) can produce images of wrong sharpness/contrast values (differing significantly from the referent one) -> false hardware/software functionality.
BBT analyzes the difference between the grabbed picture from the reference DTV/set-up-box device. Based on the expected/allowed difference it is decided whether the DUT is of acceptable functionality or not, regarding the sharpness/contrast.
Shift Errors
Errors related to shift can come from two different reasons:
- Default shift can vary in different DTV/Set-up box sets due to erroneous parameters in software -> errors in API software layer.
- Changing the shift within the device under test (DUT) can produce images of wrong shift values (differing significantly from the referent one) -> false hardware/software functionality.
BBT analyzes the difference between the grabbed picture from the reference DTV/set-up-box device. Based on the expected/allowed difference it is decided whether the DUT is of acceptable functionality or not, regarding the shift.
Menu Functionality Errors
Errors related to menu functionality in different DTV/Set-up box sets can occur due to:
- Software faults. -> errors in middleware or API software layer.
- Faults in the tuner
BBT analyzes the difference between the grabbed picture from the reference DTV/set-up-box device. Based on the expected/allowed difference it is decided whether the DUT is of acceptable functionality or not, regarding the menu.
Menu navigation Errors
Errors related to menu navigation in different DTV/Set-up box sets can occur due to software faults. -> errors in middleware or API software layer.
BBT analyzes the difference between the grabbed picture from the reference DTV/set-up-box device. Based on the expected/allowed difference it is decided whether the DUT is of acceptable functionality or not, regarding the menu navigation: the appropriate action within the DUT is initiated and the idea is to check whether the menu in the DUT has moved to appropriate position.
| < Prev | Next > |
|---|