Azure Computer Vision
An Azure artificial intelligence service that analyzes content in images and video.
379 questions
How can i change the input dimensions for spatial analysis container for processing live video feed?
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(172.8, 11%, 26%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EMH%3C/text%3E%3C/svg%3E)
Muhammad Hamad Akram
0
Reputation points
I am processing a live stream on Desktop machine by making it an iotedge device using Azure Spatial Analysis container. The size of video frame that my camera produces is 704 x 576 and the person detector used in the container is around size 416 x 416. I am wondering is there any way i can change the this and use a network that is trained of images of larger size? This might increase our accuracy.
{count} votes
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(195.2, 4%, 28%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EVM%3C/text%3E%3C/svg%3E)
VasaviLankipalle-MSFT 17,471 Reputation points2024-07-18T18:51:34.2+00:00 Hello @Muhammad Hamad Akram , Thanks for using Microsoft Q&A Platform.
Have you tried setting the
INPUT_VIDEO_WIDTHparameter? It defines the frame width of the input video/stream. This is an optional field; if specified, the frame will be scaled to this width while preserving the aspect ratio. -
Muhammad Hamad Akram 0 Reputation points
2024-07-19T11:38:36.9366667+00:00 @VasaviLankipalle-MSFT Hey Thanks for your response. I think i phrased the question differently but what I really wanted to ask was is there anyway I can specify to use a model that is trained on larger input images like a model that is trained on 640 x 640 size images or even bigger i think this might increase detection accuracy a bit. Currently spatial analysis container loads a model which takes input of size 512 X 512 something.
Sign in to comment
Sign in to answer