Original subtitles

Okay.

Then in the previous video, we have trained up to 63 epochs and got accuracy with P of 0.72.

In this video we will try to detect face mask using the train weights.

The notebook use is the same as before.

The goal of seven notebook.

If you haven't already mounted Google Drive, run the following three cells.

If so, skip ahead to the face mask detection section.

For detection.

We have provided an image that can be downloaded.

Therefore, the first code cell is used to download the image.

Use the command g down, then link the image.

When the sale by pressing the following button.

The image is facemask doc PNC.

Next move the image to the inference folder.

Use the command mv face, mask, dock and Z inference.

When the sale by pressing the following button.

Next we will detect object on the face mask, dot P and G image.

Use the following command to perform detection.

Titan detect the pine.

That's that's why it's we use what's best, not pity, which is the words that have the highest p value.

In contrast, we use 0.5.

In the image size, we use 640.

In source read inference face Mass dot three and G.

When the sale by pressing the following button.

Wait until the detection process is finished.

When finished, the detection results will be stored in the following folder.

Next, we add a function to display the image in this code cell.

This function is also used in the detection on image section.

When the sale by pressing the following button.

Continue to the next cell.

This cell will call the cell image function to display the detection result.

Blocking its path like this.

Then copy by pressing control.

See?

And there were quotes.

Place the image pad by pressing control fee.

When the sale by pressing the following button.

Here are the results.

The mask is detected properly.

Four masks are correctly detected in this image.

You can save the training results and run detection on your own computer.

To perform the detection.

You just need to download the weights with the highest MLP memory based doped.

Congratulations.

You have finished the detection project.

See you in the next video.