search

google-ai-edge / mediapipe

Cross-platform, customizable ML solutions for live and streaming media.
https://ai.google.dev/edge/mediapipe
Apache License 2.0
27.47k stars 5.15k forks source link

how can i use single image as input #12

Closed Nicol-Bolas closed 5 years ago

Nicol-Bolas commented 5 years ago

i want to test your BlazeFace on a single image, which calculator should i use? or take an opencv Mat as input of ImageTransformationCalculator?

chuoling commented 5 years ago

You could try modifying the desktop TFLite object detection example in https://github.com/google/mediapipe/blob/master/mediapipe/docs/object_detection_desktop.md#tensorflow-lite-object-detection-demo to take a single image as input, e.g., replacing OpenCvVideoDecoderCalculator with https://github.com/google/mediapipe/blob/master/mediapipe/calculators/image/opencv_encoded_image_to_image_frame_calculator.cc.

mgyong commented 5 years ago

@Nicol-Bolas Thanks for using MediaPipe. My team would love to get feedback from you about your use of MediaPipe

It would be great if you could answer the below questions

What is your use case for MediaPipe? Any feedback about how easy or difficult it is to use MediaPipe? Thanks Ming