Computer vision microsoft认知服务api中是否有名人检测的最低置信度?
我所看到的所有关于名人侦测的秘密都非常高。是否存在一些API不考虑名人的最小可信度阈值? < P>如果API不识别名人,API将返回空的或空的<代码>“名人”< /代码> JSON数组,只返回“<代码> >名人”< /代码> JSON数组,如果<代码>“信心”级别高于Computer vision microsoft认知服务api中是否有名人检测的最低置信度?,computer-vision,detection,microsoft-cognitive,Computer Vision,Detection,Microsoft Cognitive,我所看到的所有关于名人侦测的秘密都非常高。是否存在一些API不考虑名人的最小可信度阈值? < P>如果API不识别名人,API将返回空的或空的“名人”< /代码> JSON数组,只返回“ >名人”< /代码> JSON数组,如果“信心”级别高于0 下面是一个示例,通过计算机视觉API页面中的部分运行迈克尔·杰克逊的两张照片: 识别Michael Jackson,并返回一个填充的“名人”JSON数组 无法识别Michael Jackson,并返回一个空的“名人”JSON数组 是的,请检查我的更
0
下面是一个示例,通过计算机视觉API页面中的部分运行迈克尔·杰克逊的两张照片:
识别Michael Jackson,并返回一个填充的“名人”
JSON数组
无法识别Michael Jackson,并返回一个空的“名人”
JSON数组
是的,请检查我的更新答案,我包括了一个例子。第一张图片,azure 99.9%相信这是迈克尔·杰克逊,第二张图片azure不认为这是迈克尔·杰克逊。。。你把我联系在一起,根本没有什么门槛感。我已经通过API运行了20k张图片,但还没有找到一位名人的自信心低于90%。根据经验,这表明90%是临界值。您声明它高于0,您能提供文档或代码吗?我链接的示例表明,如果置信度低于0,API将返回一个空的“名人”JSON数组,因此没有置信度,是的,文档中声明“置信度,介于0和1之间”:
{
"categories": [
{
"name": "people_portrait",
"score": 0.62109375,
"detail": {
"celebrities": [
{
"name": "Michael Jackson",
"faceRectangle": {
"left": 377,
"top": 282,
"width": 381,
"height": 381
},
"confidence": 0.99902832508087158
}
],
"landmarks": null
}
}
],
"adult": null,
"tags": [
{
"name": "person",
"confidence": 0.99581664800643921
}
],
"description": {
"tags": [
"person",
"man",
"young",
"standing",
"looking",
"front",
"holding",
"wearing",
"shirt",
"posing",
"boy",
"white",
"smiling",
"dark",
"black",
"table",
"large",
"suit",
"woman",
"board"
],
"captions": [
{
"text": "Michael Jackson posing for the camera",
"confidence": 0.91584373926593021
}
]
},
"requestId": "ac41aa84-9a70-4ffc-8de4-5a662600fc56",
"metadata": {
"width": 1000,
"height": 1000,
"format": "Jpeg"
},
"faces": [
{
"age": 31,
"gender": "Male",
"faceRectangle": {
"left": 377,
"top": 282,
"width": 381,
"height": 381
}
}
],
"color": {
"dominantColorForeground": "White",
"dominantColorBackground": "Black",
"dominantColors": [
"Black",
"White"
],
"accentColor": "B64215",
"isBWImg": false
},
"imageType": {
"clipArtType": 0,
"lineDrawingType": 0
}
}
{
"categories": [
{
"name": "people_",
"score": 0.94140625,
"detail": {
"celebrities": [],
"landmarks": null
}
}
],
"adult": null,
"tags": [
{
"name": "person",
"confidence": 0.99912232160568237
},
{
"name": "outdoor",
"confidence": 0.936089813709259
}
],
"description": {
"tags": [
"person",
"outdoor",
"man",
"woman",
"standing",
"front",
"suit",
"holding",
"car",
"wearing",
"people",
"couple",
"posing",
"young",
"bus",
"black",
"smiling",
"glasses",
"umbrella",
"group",
"table",
"city",
"red",
"water",
"street",
"phone",
"boat",
"train"
],
"captions": [
{
"text": "a couple of people posing for the camera",
"confidence": 0.86922603629725859
}
]
},
"requestId": "7d57493c-f8b0-468f-8135-17da555d2463",
"metadata": {
"width": 1000,
"height": 1000,
"format": "Jpeg"
},
"faces": [
{
"age": 38,
"gender": "Female",
"faceRectangle": {
"left": 400,
"top": 270,
"width": 234,
"height": 234
}
}
],
"color": {
"dominantColorForeground": "Black",
"dominantColorBackground": "Black",
"dominantColors": [
"Black",
"Grey"
],
"accentColor": "364E5D",
"isBWImg": false
},
"imageType": {
"clipArtType": 0,
"lineDrawingType": 0
}
}