For example, in [document judgment] of this site, like the image below, it cuts out the movie captured from the camera, makes it a still image, throws it to the AI service, recognizes characters, and reads it by speech synthesis.

Other functions can also be tried quite easily, so please try it! ! Please refer to the list below as well.

Let's set the AI service like this(Qiita article will be displayed) after permission of camera / microphone. Be careful not to add unnecessary spaces or "/" at the end of the URL.

Also, it will not work if the expiration date of the key (Microsoft Cognitive Services is about 30 days from acquisition) has expired. Please check at the site of each service.

If that is not the case, try clearing your browser's cache and rereading and trying the settings again.

The control panel is displayed on the right side of the screen. From the top, it becomes "camera information", "microphone information", "message" displaying speech synthesis messages, etc., operation panel sending commands to the Smart Agent, "developer console" displaying debugging information.

The [try] button allows you to perform each function using AI. Explain button is explained by voice. Since I have separated the long explanation, please push it again when the explanation is over. (The 【】 icon in each heading of "Q&A" is also the same)

If speech recognition is effective, pages are automatically moved by telling the title of the page, such as "Production company" and "What is this".

If you get strange behavior, please reload the browser. Processing is reset and it returns to the initial state. Speech synthesis is also canceled.

After that, please press the [next] button on the "Try Simple AI" page, approve the camera and microphone, and try the respective functions such as [face image judgment].

Even if speech recognition is enabled, it is canceled, so please be careful. To activate speech recognition, click the [try again] button with [speech recognition] again.

No relation at all. "Simple AI" is manufactured independently by Denaripam Co., Ltd.

So, even if you ask AI service provider about "Simple AI", we will not be able to answer from the other side. If there is something, please listen from the "Feedback" on this page! !

It does not, it completely works with the client (browser) only. Communication is AI service only.

The control panel screen on the right side also seems to be linked with the server, but I have not done any. Just cut out the image and audio data to the AI service and pass it only when you try the AI service with buttons or instructions.

Although it has merit and restriction of ...... to operate with javascript, since it operates only with the browser, I think that smart speaker with display etc can be created in combination with tablet.

Also, since it is relatively easy to recognize the face, it can be used as a data generation tool that hands over existing camera images and collects age gender information.

If you have any good ideas, I would be happy if you give us your opinion!

For example, by adding the following code to index.html included in the downloaded framework, you can call youtube with speech recognition.

If your company has a publicity channel on youtube, you can send a publicity movie with speech recognition by telling the company name.

//↓Insert in downloaded line 276 of index.html------------------------------
			{
				label: "SimpleAI.productcall-start",
				word:[/##Add a company name##/],	//Register company name here # is unnecessary
//				word:[/Denaripam/],			//Sample
				messageafter:function(sender,event,param) {		//Processing to be executed after message playback upon recognition.javascript.
					$("#informationmodal").modal();
					YouTubeplayer_start("##Enter company publicity keywords##");	//Submit youtube public relations video keyword xxx official etc.
				},
			},
//↑------------------------------

//↓Insert in downloaded line 115 of index.html------------------------------
var YouTubeplayer;
var YouTubeIframeAPI=false;
 function onYouTubeIframeAPIReady() {
	YouTubeIframeAPI=true;
}
var YouTubeplayer_start=function(keyword,volume) {
	if(YouTubeIframeAPI) {
		YouTubeplayer = new YT.Player("player", {
			height: "360",
			width: "640",
			events: {
				onReady: function(event){
					if(keyword) YouTubeplayer.loadPlaylist({listType:"search",list:keyword,});
					if(volume)  YouTubeplayer.setVolume(100);
				},
			},
		});
	}
};
$(window).on("load", function(){
	//informationmodal
	$("#informationmodal").on("hidden.bs.modal", function () {
		if(YouTubeplayer)YouTubeplayer.destroy();
	});
});
/↑------------------------------

//↓Insert in downloaded line 112 of index.html------------------------------
<script type="text/javascript" src="https://www.youtube.com/iframe_api"></script>
//↑------------------------------

Speech recognition is adjusted with a directional microphone (made by Realtek) of a delicate low price laptop computer.

Since we are doing voice collection etc. with soft approach, it seems that it depends considerably on terminal to use.

The settings on this site are embedded so they can not be changed, but in the download file you can adjust with the following code.

//↓Insert in downloaded line 162 of index.html------------------------------
		MicrophoneRecordparams:{
			// Start recording when exceeding the threshold,% specified 1-99  n			autostartRecordingRangePer:10,
			// If a sound is generated within the recording end detection period, the value considered as continuation of recording (relaxation of unintended speech delimitation). % Specified 1-99  n 			autostopRecordingRangePer:10,
			// silence detection period, delimit sound if no sound beyond autostopRecordingRangePer within this period ms  n			autostopRecordingCheckTime:700,
// Minimum length of the sound. If it is not longer than this, it is not subject to speech. Because we do not wastefully consume voice recognition transaction. ms  n 			MinChunksTime:1000,	
		},
//↑------------------------------

You can disable speech synthesis in "Mute mode". Speech recognition can be canceled with "Speech recognition cancellation". Resume speech synthesis again in "mute mode", resume speech recognition with "speech recognition execution". (Please send a message from the operation panel on the right)

※ The following functions are noisy so they were disabled by default.

If speech synthesis is enabled and speech recognition is disabled, if the microphone receives a somewhat long speech, it will be said that "speech recognition is canceled". (Cough, sneeze etc)

When speech synthesis is enabled and speech recognition is also effective, when the microphone receives a sound that can be recognized as a word, it is handed over to the Smart Agent, so if there is no such command, "I am sorry, again Please be uttered "Please.

Basic There is no guarantee and no support, but the production company is doing support etc. for paid.

"Simple AI" Overview

"Simple AI" the purpose

Overview of application product

About other efforts for AI