Preprocessing images to make them suitable for a trained AI model is an important step as it improves the accuracy of image inferencing.
This webinar will explore:
- various image preprocessing requirements for AI inference and how they influence the accuracy of a trained model
- some typical image preprocessing techniques, such as camera image signal processing, format conversion, mean subtraction and normalization
- image processing libraries, such as OpenCV and the pipeline based multimedia framework "Gstreamer"
- preprocessing requirements for different model frameworks, such as Caffe and TensorFlow, when using models from a Model Zoo.
At the end of the presentation you will also see a practical example of configuring image preprocessing for an RZ-V2L microprocessor from the RZ Family of MPUs from Renesas.
Rahul Dubey - Doulos Member Technical Staff will be presenting this training webinar, which will consist of a one-hour session (see below for details) and be interactive with Q&A participation from attendees.
Attendance is free of charge
If you have any queries, please contact firstname.lastname@example.org
More FREE online support resources with Doulos KnowHow
Doulos Knowhow includes
- Technical information
- Coding examples
- Video guides
AI, Deep Learning & Python training available now from Doulos:
Please contact Doulos to schedule a public course to suit you, or to discuss onsite training.