I am using TensorFlow.js in my web application and the input is a camera. In particular, I am using MobileNet to classifying the camera input. This tutorial is an inspiration for this app. There is a video object:
<video autoplay playsinline muted id="webcam" width="224" height="224"></video>
And I access it in my JavaScript file:
const webcamElement = document.getElementById('webcam');
And take an image from the video object, convert it into a tensor to later preprocess and predict using the frame:
const webcam = await tf.data.webcam(webcamElement);
However, when I run the last line above, it results in the following error:
Am I doing something wrong? I tried to run the tutorial which also used tf.data.webcam()
and it worked ok.
This error disappeared when I restarted my computer and reloaded the webpage.