Stack Overflow
Stack Overflow

Questions tagged [mediapipe]

When you are facing any implementation issue for mediapipe in your android, iOS or desktop application, When you want any clarification on concept , methods on mediapipe etc ask question with this tag.

MediaPipe is a framework for building multimodal (eg. video, audio, any time series data), cross platform (i.e Android, iOS, web, edge devices) applied ML pipelines. With MediaPipe, a perception pipeline can be built as a graph of modular components, including, for instance, inference models (e.g., TensorFlow, TFLite) and media processing functions.

585 questions
1
vote
0 answers

Compile pbtxt into binarypb

I'm playing around with Mediapipe and I'm trying to better understand how the graph works and what is the input/output of the different calculators. If I understand correctly, the .pbtxt files are just plain-text instructions that describe how each…
alexandernst
  • 14,352
  • 22
  • 97
  • 197
1
vote
1 answer

How to use 2 mediapipe models using react

Im trying to parse video using 2 mediapipe models (FaceMesh and FaceDetection) however, I cannot find a way to draw landmarks from both models. I can see in console outputs from both console.log(results) but I can only see landmarks from model which…
Roitko
  • 117
  • 1
  • 8
1
vote
1 answer

How to calculate gaze position?

I am trying to code out an eye tracker using Python. I am using the Face Mesh solution from the Mediapipe library to track the face, and I am able to get the x and y coordinates of the center of both irises. import cv2 import mediapipe as mp import…
nicesaj360
  • 21
  • 2
1
vote
0 answers

Can't use high version of mediapipe on old python version

I am on conda venv with python version 3.6.13, also on mediapipe version 0.8.3, when running a python script (below) on a higher version of mediapipe version 0.8.11 and code crashes. NOTE: I tried to install mediapipe version 0.8.11 on conda venv…
alibustami
  • 198
  • 2
  • 13
1
vote
0 answers

Segmentation fault after installing mediapipe on raspberry pi 3 model b

After updating numpy to its latest version I tried to install mediapipe on my raspberry pi 3 model b with following instructions found in here... sudo apt install ffmpeg python3-opencv python3-pip sudo apt install libxcb-shm0 libcdio-paranoia-dev…
Seyed Kian
  • 45
  • 5
1
vote
0 answers

How to change colors of the tracking points and connector lines on the output using Mediapipe Pose for Android?

I am referring to 33 body points and connector lines between them. I'd like to change the colors of those, especially of the white default color of the connector lines. Here's my code: @Override protected void onCreate(Bundle savedInstanceState) { …
Vishal Bhosle
  • 33
  • 3
1
vote
0 answers

How do i generate a dll file for Selfie Segmentation Mediapipe

How do i go about generating a dll file for Selfie Segmentation Mediapipe. I just couldn't figure it out since there is no source code specifically for Selfie Segmentation part, do i need to call my own functions? or is there any other way doing…
Kishore
  • 11
  • 1
1
vote
0 answers

Is there any good Mediapipe documentation?

i would like to learn to work with mediapipe. Are there any good documentations for this framework ? I would be grateful for any help, becausde i find the official documentation no really usefull. Greets
Daniel
  • 163
  • 12
1
vote
0 answers

How to run mediapipe hand detection to Coral usb accelerator

Is it possible to run mediapipe hand detection with Coral usb accelerator? I have seen plenty of tutorial about tflite but nothing about mediapipe
Liampour
  • 51
  • 1
  • 5
1
vote
1 answer

Merging Errors: Attribute meta-data requires a placeholder substitution but no value is provided. BazelFaceMesh.app main manifest (this file)

I am migrating the Mediapipe face mesh app from Bazel to Gradle, but when building the app it gives me the following error: Merging Errors: Error: Attribute meta-data#cameraFacingFront@value at AndroidManifest.xml:38:53-89 requires a placeholder…
Lord Konzi
  • 11
  • 1
1
vote
1 answer

mediapipe.mp_holistic 'module' object is not callable

def PoseEstimation(frame,win_name): with mp_holistic(min_detection_confidence=0.5,min_tracking_confidence=0.5) as holistic frame = cv.cvtColor(frame, cv.COLOR_BGR2RGB) results = holistic.process(frame) frame =…
Fish Ung
  • 17
  • 8
1
vote
0 answers

How to use mediapipe face mesh on pictures

I modified the official example and tried to put the face mesh on the picture https://google.github.io/mediapipe/solutions/face_mesh.html#javascript-solution-api const faceMesh = new FaceMesh({locateFile: (file) => { return…
Zac
  • 11
  • 3
1
vote
0 answers

AttributeError: module 'mediapipe' has no attribute 'solutions'

I am running the media pipe face detection, when I run this binary given in media pipe documentation, I followed the same steps as given, but it's giving the following error. AttributeError: module 'media pipe' has no attribute 'solutions' import…
user15320062
1
vote
1 answer

Modify Mediapipe C++ Hand Detection example

I'm trying to extract landmarks from Hand Detection graph while modifying demo_run_graph_main.cc. To do so, i modified hand_tracking_desktop_live.pbtxt adding following line: output_stream: "landmarks" I did that because in node…
Mikhail
  • 48
  • 5
1
vote
0 answers

How to retrieve confidence scores for pose landmarks?

I'm applying MediaPipe Holistic to scenes that can be challenging to parse. I would like to retrieve in Python the computed confidence score for each landmark returned by MediaPipe's pose model. After a few hours of searching the answer seems to be…
Ronnen Levinson
  • 133
  • 7
1 2 3
…
38 39