How To Add Code to Upload image and show image that has keypoint from YoloV8 model?

Question

I have run the YoloV8 model example for estimating animal poses on Google Colab and I want to add code to upload animal images and be able to display image results in the form of animal images that already have keypoints with the model that has been created on https://colab.research. google.com/drive/12YNRtDUWpX95Nh3rd9gMvCZnjN8jenor?usp=sharing.

I get the code from https://github.com/thomasreynolds4881/Animal-Keypoint-Estimation/blob/master/estimation.ipynb

Can someone help me make the code I mean? Pleasee

From the very end of my Google Colab, I tried to create code to upload images and it failed, it only displayed the animal image that I had just uploaded.

What I want is, the image that has just been uploaded will immediately appear in the form of an animal image output that already has keypoint annotations from the model that has been created

I hope someone can help me, by writing code to create an image upload function and displaying the processed output image with keypoint annotations

Answer 1

I recently encountered a similar problem on training a keypoint pose estimator on an animal dataset using YOLOv8. The issue was that I was missing the save=True argument when I ran an inference on an image. Try running this, replacing the path:

results = model('/path/to/input_image.jpg', save=True)

The result should save to a directory called /predict. You can either navigate to this directory manually, or display it with this:

import matplotlib.pyplot as plt
import matplotlib.image as mpimg

img_path = '/runs/pose/predict/input_image.jpg' #replace with your path
img = mpimg.imread(img_path)
plt.imshow(img)