Network-type for secondary pgie

[edit_or]

I have a few queries.

(1)My secondary pgie is a sort of text reading model using CTCGreedyDecoder.
To decode text reading output from output layer of secondary model, do I need raw-output-generated-callback function to implement the result?

(2)Secondary pgie has fixed input size 24,94.
The following parameters are set in config file.
input-object-min-width=94
input-object-min-height=24
input-object-max-width=94
input-object-max-height=24
Will it resize the input to secondary pgie?

(3)Primary pgie has two classes in its detection.
Secondary pgie uses only second class, so operate-on-class-ids=2 is set in configuration file. Is it enough?

(4)If I need to modify a bit on bounding boxes between primary pgie and secondary pgie, I can use plugin to extract the boxes. But how to put it back as input to secondary pgie?

The issue is that output has display only for primary pgie. Secondary pgie doesn’t show output at display yet.
How can I debug?

[edit_or]

I found here that input to sgie is cropped, and resized.

SGIE will crop the object from NvStreamMux buffer using the object's bbox detected by the Primary GIE. The crop is then scaled/converted to the network resolution/color format. For example, if the NvStreamMux resolution is 1920x1080, SGIE will crop using object bbox co-ordinates (e.g. x=1000, y=20, w=400, y=500) from the 1920x1080 image and then scale it to the SGIE network resolution (say 224x224). In practice, the object crop + scaling + color conversion happens in one go.