"Checks" section#

The "Checks" section allows you to upload media files and evaluate them using the selected checks (Figure 94).

Checks are available only if the user has the corresponding licenses.

Image check#

To start evaluating a photo:

Select the "Image" block;
Upload or drag your file;
Select checks from the list:
- Liveness and image modification;
- DeepFake and image modification;
- Compliance with ISO/IEC 19794-5:2011;
- Compliance with ICAO standard;
- Compliance with with the biometric standards;
Start the check.

The results page contains (Figure 95):

Final assessment of whether the photo passed the selected checks;
Results for each check criterion: if the photo meets the criterion, the rating color is green is green; if not, the color is red. A photo must meet all the criterion in order to comply with the check;
The number of criteria that the photo does not meet if the check result is negative;
A button to download the check results in JSON;
A button to reset old photo . Allows you to start a new check for a different photo;
A button for selecting checks from the list of available ones .

Figure 94. Figure 95. Photo evaluation results

File requirements:

.jpeg, .png, .tif, or .bmp format;
Size no more than 15 MB and no more than 3840×2160 pixels;
The image may contain one or more people;
Image must have a person's face.

Video check#

To start evaluating a video file:

Select the video block;
Upload or drag your file;
Select checks from the list:
- Liveness;
- DeepFake;
- Face attributes;
- Body attributes;
Set video processing parameters if needed via the " Settings" button (Table 35);
Start the check.

Advanced settings

Advanced settings

To ensure correct evaluation results, it is not recommended to change these parameters. If necessary, reset the parameters to their default values using the button at the bottom of the window.

Table 35. Check configuration

Parameter	Description
Video processing parameters
Frame analysis frequency measurement type	Frame rate unit: seconds or frames
Frame analysis frequency	Depending on the selected frame analysis frequency measurement type: The interval in seconds after which a frame is selected for analysis. By default, frames are processed every 10 seconds; The interval between frames that will be selected for analysis: the first and every N+1 frame will be taken, where N is the specified value. By default, every 10 frames are processedю The selected frames are then analyzed and, if the threshold value is met, are selected to create a track
Maximum number of frames used for face attribute estimation	Maximum number of best frames in a track used to evaluate face attributes
Number of consecutive frames with a face	Allows you to specify the number of frames required to create a track. The track begins when a detection with face or body appears on the last frame of the specified number of consecutive frames
Filters
Reject detections with quality below	Ignore low-quality detections
Discard face images with head rotation angle (to the left or right, yaw) above	Ignoring images in which the person's head is turned to the left or right at a too large angle —no information will be extracted when detecting a face and evaluating the angle of head rotation.
Discard face images with head tilt angle (to the left or right, roll) above	Ignoring images in which a person's head is tilted to the left or right at a too large angle—no information will be extracted during face detection and head tilt evaluating.
Discard face images with head tilt angle (up or down, pitch) above	Ignoring images in which the person’s head is tilted up or down at a too large angle —no information will be extracted during face detection and head tilt evaluating.

File requirements:

.avi, .mp4, .mov, .mkv, .flv, .quicktime fprmat;
File size depends on the specified timeout during which the connection with the service is established — up to 2 GB. For details, see the LUNA_VIDEO_AGENT_TIMEOUTS parameters in the LUNA PLATFORM 5 documentation

The results page contains (Figure 96):

A window with the processed video. Recognized objects, bodies, and faces are bounded in bboxes;
A time track with per-second marking indicating the current video time and evaluation results. Move the pointer to see frame-by-frame results. From the list of detected persons in the frame, select a face or body to view the evaluation for that subject;
Time bars with color-coded results per segment of the video depending on the selected checks. For Liveness and DeepFake:
- Live/real face—green segment;
- Fake face—red segment;
- Unknown—gray segment;
Results for each check: if the video passes the check, the final score is green; if it fails, it is red.
A button to reset old video . Allows you to start a new check for a different file;
A button for selecting verifications from the list of available ones .

Figure 94. Figure 96. Video evaluation results

Audio and/or video evaluation (beta)#

Audio analysis for Deepfake detection is performed by the Audio Spoofing module for LUNA PLATFORM 5.

For files containing audio and/or video, Deepfake evaluation is available for fake faces and/or voices. To start:

Select the video and/or audio block";
Upload or drag the file;
Start the evaluation.

File requirements:

.mp4, .mp3, .ogg, .aac
Maximum file size—500 MB

The results page contains the same blocks as video evaluation. Move the pointer to see frame-by-frame or per-second results. Time bars show evaluation results with color coding by segment. For Deepfake results:

Real face and/or voice—green segment;
Fake face and/or voice—red segment;
Unknown—gray segment.

Click the reset button next to the current file to start a new evaluation with a different file.