Hello,
I have a 1280x800 image, and I'm trying to resize it with stretching using an ImageManip node to 300x300. My goal is to send the resized image to MobileNetSpatialDetectionNetwork.
Instead of stretching the image, it resizes and crops is vertically. Here's what I mean:
Input image: (ignore grayscale/green lines)
Expected:
Actual:
You can see the actual is cropped at the top instead of being stretched. Here is the image manip code:
resizeNode = pipeline.create(dai.node.ImageManip)
resizeNode.initialConfig.setResize(300, 300)
Question 1: What config do I need to set to have it stretch the image instead of cropping?
Question 2: Is there an easier way to stretch the image for using with a neural network node?
Question 3: Does the spatial neural network node expect the stretched image? Does it automatically map it to the depth image correctly? The input depth & color images are perfectly aligned.
