edge-impulse-classify.html

<script type="text/javascript">
    RED.nodes.registerType('edge-impulse-classify',{
        category: 'function',
        color: '#FFCC66',
        defaults: {
            name: {value:""},
            path: {value:""}
        },
        inputs:1,
        outputs:1,
        icon: "edge-impulse-logo.png",
        label: function() {
            return this.name||"edge impulse classify";
        }
    });
</script>

<script type="text/html" data-template-name="edge-impulse-classify">
    <div class="form-row">
        <label for="node-input-name"><i class="fa fa-tag"></i> Name</label>
        <input type="text" id="node-input-name" placeholder="Name">
    </div>
    <div class="form-row">
        <label for="node-input-path"><i class="fa fa-tag"></i> Path</label>
        <input type="text" id="node-input-path">
    </div>
</script>

<script type="text/html" data-help-name="edge-impulse-classify">
    <p>Performs an "edge impulse" classification and/or anomaly detection.</p>
    <h3>Inputs</h3>
    <dl class="message-properties">
        <dt>Path <span class="property-type">configuration property</span>
        </dt>
        <dd><p>This is the folder where you have deployed the "edge impulse" WebAssembly library (see Prerequisites).</p>
            <p>So it is the folder where your Node-RED application should find the 2 files: <code>edge-impulse-standalone.js</code> and <code>edge-impulse-standalone.wasm</code></p>
        </dd>
        <dt>payload
            <span class="property-type">array</span>
        </dt>
        <dd><p>The input array that should be classified.</p>
            <p>In case of audio classification it is typically the raw audio format (PCM) 
            of the audio fragment that should be classified converted into a javascript array.  
            It is important that this array has exactly the same size as the size of the audio fragments that are used for training the edge impulse.</p>
        </dd>
    </dl>
    <h3>Outputs</h3>
    <ol class="node-ports">
        <li>Standard output
            <dl class="message-properties">
                <dt>payload <span class="property-type">Object</span></dt>
                <dd>the classification outcome.  An example of such an outcome:
                    <pre>{ "anomaly":0,
  "results":[{"label":"other","value":1},
             {"label":"ring", "value":2.582810076923181e-10}]}</pre>
                </dd>
            </dl>
        </li>
    </ol>
    <h3>Prerequisites</h3>
    <p>The "edge impulse", you (or someone else) have created and trained at <a href="https://www.edgeimpulse.com/">edge impulse</a>, 
       should be deployed as a WebAssembly library to a folder readable by your Node-RED application.
    <p>It is easy to download your edge impulse project as a WebAssembly library using the online edge impulse studio: for this you need to select the "Deployment" menu, 
       then select create "WebAssembly" library and then select "Build".  This will download a zip file consisting of 2 files: <code>edge-impulse-standalone.js</code> and <code>edge-impulse-standalone.wasm</code>.  
       So these 2 files must be copied to a folder that is accessible by your Node-RED application.
    </p>
    <h3>Details</h3>
    For more details consult the <a href="https://github.com/janvda/node-red-contrib-edge-impulse">GitHub Readme</a> 
    <h3>References</h3>
    <ul>
        <li><a href="https://www.edgeimpulse.com/">Edge Impulse</a> - is the leading development platform for machine learning on edge devices, free for developers and trusted by enterprises.</li>
        <li><a href="https://github.com/janvda/node-red-contrib-edge-impulse">GitHub janvda/node-red-contrib-edge-impulse</a> - the github repository for this node.  Feel free to open new issues.</li>
        <li><a href="https://github.com/janvda/node-red-doorbell">GitHub janvda/node-red-doorbell</a> - a project using this node to recognize the ring sound of a doorbell.</li> 
    </ul>
</script>