You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/docs/Markup-Guide.md
+4-4Lines changed: 4 additions & 4 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -16,7 +16,7 @@ default values can be changed be setting the system property listed for each:
16
16
17
17
-`MARKUP_LABELS_ENABLED`
18
18
- System property: `markup.labels.enabled`
19
-
- Default value: `true`
19
+
- Default value: `true`
20
20
- If true, add a label to each detection box.
21
21
-`MARKUP_LABELS_ALPHA`
22
22
- System property: `markup.labels.alpha`
@@ -80,7 +80,7 @@ default values can be changed be setting the system property listed for each:
80
80
- The CRF value can be from 0-63. Lower values mean better quality. Recommended values range from 15-35, with 31 being recommended for 1080p HD video. This property is only used if generating VP9-encoded `.webm` files
81
81
-`MARKUP_ANIMATION_ENABLED`
82
82
- System property: `markup.video.animation.enabled`
83
-
- Default value: `true`
83
+
- Default value: `false`
84
84
- If true, draw bounding boxes to fill in the gaps between detections in each track. Interpolate size and position.
85
85
86
86
# Video Markup Icons
@@ -115,10 +115,10 @@ The frame above shows a paper clip icon to indicate that the detection is the re
115
115

116
116
117
117
The frame above shows the person detection in addition to a new skis detection. The confidence for the latter is lower, which is good considering the algorithm misclassified the person's shadow as skis. The skis track is only a few frames long, so the WFM determined it was a non-moving (stationary) track. This is represented by the anchor icon at the start of the label. Also, notice that the labels are semi-transparent. This allows you to read labels and see frame content that would otherwise be hidden if the labels were completely opaque. Note that you may want to set `MARKUP_LABELS_ALPHA` to `0.75` or greater when using the `mjpeg` encoder.
118
-
118
+
119
119
# Video Encoder Considerations
120
120
121
-
Performing markup on an image will always generate a `.png` file. Performing markup on a video will generate a video file based on the value of `MARKUP_VIDEO_ENCODER`. The `vp9`, `h264`, and `mjpg` encoders are supported.
121
+
Performing markup on an image will always generate a `.png` file. Performing markup on a video will generate a video file based on the value of `MARKUP_VIDEO_ENCODER`. The `vp9`, `h264`, and `mjpg` encoders are supported.
122
122
123
123
The `vp9` and `h264` encoders serve the same purpose in that both formats can be played in the WFM web UI in most web browsers, while the `.avi` files resulting from the `mjpeg` format must be downloaded and played using a separate program like [VLC](https://www.videolan.org/vlc/index.html) or [mpv](https://mpv.io/). In general, `h264` encoding is much faster than `vp9` encoding, so you may want to use it instead of `vp9`. Please be aware that you may be required to pay [Usage Royalties](License-And-Distribution.md#usage-royalties) when using the `h264` encoder for commercial purposes.
<p><imgalt="Walking With Border Skies" src="../img/markup/walking-with-border-skis.jpg" title="Walking With Border Skis" /></p>
406
406
<p>The frame above shows the person detection in addition to a new skis detection. The confidence for the latter is lower, which is good considering the algorithm misclassified the person's shadow as skis. The skis track is only a few frames long, so the WFM determined it was a non-moving (stationary) track. This is represented by the anchor icon at the start of the label. Also, notice that the labels are semi-transparent. This allows you to read labels and see frame content that would otherwise be hidden if the labels were completely opaque. Note that you may want to set <code>MARKUP_LABELS_ALPHA</code> to <code>0.75</code> or greater when using the <code>mjpeg</code> encoder.</p>
<p>Performing markup on an image will always generate a <code>.png</code> file. Performing markup on a video will generate a video file based on the value of <code>MARKUP_VIDEO_ENCODER</code>. The <code>vp9</code>, <code>h264</code>, and <code>mjpg</code> encoders are supported.</p>
408
+
<p>Performing markup on an image will always generate a <code>.png</code> file. Performing markup on a video will generate a video file based on the value of <code>MARKUP_VIDEO_ENCODER</code>. The <code>vp9</code>, <code>h264</code>, and <code>mjpg</code> encoders are supported.</p>
409
409
<p>The <code>vp9</code> and <code>h264</code> encoders serve the same purpose in that both formats can be played in the WFM web UI in most web browsers, while the <code>.avi</code> files resulting from the <code>mjpeg</code> format must be downloaded and played using a separate program like <ahref="https://www.videolan.org/vlc/index.html">VLC</a> or <ahref="https://mpv.io/">mpv</a>. In general, <code>h264</code> encoding is much faster than <code>vp9</code> encoding, so you may want to use it instead of <code>vp9</code>. Please be aware that you may be required to pay <ahref="../License-And-Distribution/index.html#usage-royalties">Usage Royalties</a> when using the <code>h264</code> encoder for commercial purposes.</p>
410
410
<p>The <code>mjpeg</code> encoder is faster than both the <code>vp9</code> and <code>h264</code> encoders, but produces lower quality video. Specifically, the label text is not as clear. You may want to use it when developing components or marking up large video files.</p>
411
411
<p>To give you a sense of performance, here are the results of a very limited batch of tests. Note that if you choose to use the <code>vp9</code> encoder, you can increase the CRF value to reduce processing time at the cost of reduced video quality.</p>
0 commit comments