Custom DNN
Example of modified DNN module with custom post-processing.
By Laurent Ittiitti@usc.eduhttp://jevois.orgGPL v3
 Language: C++Supports mappings with USB output: YesSupports mappings with NO USB output: Yes 
 Video Mapping:   JVUI 0 0 30.0 CropScale=RGB24@1024x576:YUYV 1920 1080 30.0 JeVois CustomDNN

Module Documentation

This example shows you how to customize the JeVois DNN processing framework. Here, we create a custom post-processor that enhances the functionality of PostProcessorDetect by extracting regions of interest for each detected object. For this example, we simply compute an edge map for each detection box and display it as an overlay. Other processing is possible, for example, detected regions of interest could be sent to another neural network for additional processing.

The main goal of this module is as a tutorial for how to implement custom operations within the JeVois DNN framework in C++. The main idea is to create derived classes over Pipeline and PostProcessorDetect.

ParameterTypeDescriptionDefaultValid Values
This module exposes no parameter
params.cfg file
# select a detection model
zoo=models.yml
pipe=NPU:Detect:yolo11n-512x288
Detailed docs:CustomDNN
Copyright:Copyright (C) 2024 by Laurent Itti, iLab and the University of Southern California
License:GPL v3
Distribution:Unrestricted
Restrictions:None
Support URL:http://jevois.org/doc
Other URL:http://iLab.usc.edu
Address:University of Southern California, HNB-07A, 3641 Watt Way, Los Angeles, CA 90089-2520, USA