> For the complete documentation index, see [llms.txt](https://doc.crisen.org/ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://doc.crisen.org/ai/tencent/image.md).

# 图片识别

## 看图说话

```
$res = $ai->vision()->path('path/to/some_image.jpg')->imgToText([
    'session_id'  => time()
]);

$res = $ai->vision()->url('http://domain.com/some_image.jpg')->imgToText([
    'session_id'  => time()
]);


if($res->success()){
    var_dump($res);
}
```

## 多标签识别

```
$res = $ai->image()->path('path/to/some_image.jpg')->tag();

if($res->success()){
    var_dump($res);
}
```

## 模糊图片识别

```
$res = $ai->image()->path('path/to/some_image.jpg')->fuzzy();

if($res->success()){
    var_dump($res);
}
```

## 美食图片识别

```
$res = $ai->image()->path('path/to/some_image.jpg')->food();

if($res->success()){
    var_dump($res);
}
```

## 场景识别

```
$res = $ai->vision()->path('path/to/some_image.jpg')->scener([
    'format'  => 1, // 图片格式 1表示JPG格式（image/jpeg）
    'topk'    => 1, // 返回结果个数
]);

if($res->success()){
    var_dump($res);
}
```

## 物体识别

```
$res = $ai->vision()->path('path/to/some_image.jpg')->object([
    'format'  => 1, // 图片格式 1表示JPG格式（image/jpeg）
    'topk'    => 1, // 返回结果个数
]);

if($res->success()){
    var_dump($res);
}
```