이미지 태그 지정(버전 4.0)

이미지 분석은 이미지에 나타나는 수천 개의 인식 가능한 개체, 생명체, 풍경 및 동작에 대한 콘텐츠 태그를 반환할 수 있습니다. 태그 지정은 전면에 있는 사람과 같은 주요 주체로 국한되지 않고, 환경(실내 또는 실외), 가구, 도구, 식물, 동물, 액세서리, 가젯 등을 포함합니다. 태그는 분류법으로 구성되지 않으며 상속 계층 구조도 없습니다. 태그가 모호하거나 누구나 알 수 있는 것이 아닌 경우 API 응답은 알려진 설정의 컨텍스트에서 태그의 의미를 명확히 설명하는 힌트를 제공합니다.

Vision Studio를 사용하여 브라우저에서 이미지 태그 지정 기능을 빠르고 쉽게 사용해 보세요.

이미지 태그 지정 예제

다음 JSON 응답은 이미지 예에서 검색된 시각적 기능에 태그를 지정할 때 Azure AI Vision이 반환하는 내용을 보여 줍니다.

A blue house and the front yard.

{
    "metadata":
    {
        "width": 300,
        "height": 200
    },
    "tagsResult":
    {
        "values":
        [
            {
                "name": "grass",
                "confidence": 0.9960499405860901
            },
            {
                "name": "outdoor",
                "confidence": 0.9956876635551453
            },
            {
                "name": "building",
                "confidence": 0.9893627166748047
            },
            {
                "name": "property",
                "confidence": 0.9853052496910095
            },
            {
                "name": "plant",
                "confidence": 0.9791355729103088
            },
            {
                "name": "sky",
                "confidence": 0.976455569267273
            },
            {
                "name": "home",
                "confidence": 0.9732913374900818
            },
            {
                "name": "house",
                "confidence": 0.9726771116256714
            },
            {
                "name": "real estate",
                "confidence": 0.972320556640625
            },
            {
                "name": "yard",
                "confidence": 0.9480281472206116
            },
            {
                "name": "siding",
                "confidence": 0.945357620716095
            },
            {
                "name": "porch",
                "confidence": 0.9410697221755981
            },
            {
                "name": "cottage",
                "confidence": 0.9143695831298828
            },
            {
                "name": "tree",
                "confidence": 0.9111745357513428
            },
            {
                "name": "farmhouse",
                "confidence": 0.8988940119743347
            },
            {
                "name": "window",
                "confidence": 0.894851803779602
            },
            {
                "name": "lawn",
                "confidence": 0.894050121307373
            },
            {
                "name": "backyard",
                "confidence": 0.8931854963302612
            },
            {
                "name": "garden buildings",
                "confidence": 0.8859137296676636
            },
            {
                "name": "roof",
                "confidence": 0.8695330619812012
            },
            {
                "name": "driveway",
                "confidence": 0.8670969009399414
            },
            {
                "name": "land lot",
                "confidence": 0.856428861618042
            },
            {
                "name": "landscaping",
                "confidence": 0.8540748357772827
            }
        ]
    }
}

API 사용

태그 지정 기능은 Analyze Image API의 일부입니다. REST를 사용하여 이 API를 호출할 수 있습니다. features 쿼리 매개 변수에 Tags를 포함합니다. 그런 다음, 전체 JSON 응답을 받으면 "tags" 섹션의 콘텐츠에 대한 문자열을 구문 분석합니다.

다음 단계