Class: VisualRecognitionV3

VisualRecognitionV3

watson.visual_recognition({version: 'v3', ...});

Parameters:
Name Type Description
options
Source:

Methods

classify(callback) → {ReadableStream|undefined}

Accepts either a url, a single image file, or a zip file with multiple images (.jpeg, .png, .gif) and scores every available classifier on each image. It then applies a threshold and returns the list of relevant classifier scores for each image. Example response: { "images": [ { "classifiers": [ { "classes": [ { "class": "animal", "score": 0.998771, "type_hierarchy": "/animals" }, { "class": "mammal", "score": 0.998499, "type_hierarchy": "/animals/mammal" }, { "class": "dog", "score": 0.900249, "type_hierarchy": "/animals/pets/dog" }, { "class": "puppy", "score": 0.5, "type_hierarchy": "/animals/pets/puppy" } ], "classifier_id": "default", "name": "default" } ], "image": "dog.jpg" } ], "images_processed": 1 }
Parameters:
Name Type Attributes Default Description
params.images_file ReadStream <optional>
 The image file (.jpg, .png, .gif) or compressed (.zip) file of images to classify. The total number of images is limited to 100. Either images_file or url must be specified.
params.url String <optional>
 The URL of an image (.jpg, .png, .gif). Redirects are followed, so you can use shortened URLs. The resolved URL is returned in the response. Either images_file or url must be specified.
params.classifier_ids Array <optional>
 ['default'] An array of classifier IDs to classify the images against.
params.owners Array <optional>
 ['me','IBM'] An array with the value(s) "IBM" and/or "me" to specify which classifiers to run.
params.threshold Number <optional>
 A floating point value that specifies the minimum score a class must have to be displayed in the response.
callback function
Source:
Returns:
Type
ReadableStream | undefined

createClassifier(params) → {ReadableStream|undefined}

Train a new classifier from example images which are uploaded. This call returns before training has completed. You'll need to use the getClassifer method to make sure the classifier has completed training and was successful before you can classify any images with the newly created classifier. Example inputs: { foo_positive_examples: fs.createReadStream('./foo-pics.zip'), negative_examples: fs.createReadStream('./not-foo-pics.zip'), name: 'to-foo-or-not' } { foo_positive_examples: fs.createReadStream('./foo-pics.zip'), bar_positive_examples: fs.createReadStream('./bar-pics.zip'), name: 'foo-vs-bar' } { foo_positive_examples: fs.createReadStream('./foo-pics.zip'), bar_positive_examples: fs.createReadStream('./bar-pics.zip'), negative_examples: fs.createReadStream('./not-foo-pics.zip'), name: 'foo-bar-not' } Example output: { "classifier_id": "fruit_679357912", "name": "fruit", "owner": "a3a48ea7-492b-448b-87d7-9dade8bde5a9", "status": "training", "created": "2016-05-23T21:50:41.680Z", "classes": [ {"class": "banana"}, {"class": "apple"} ] }
Parameters:
Name Type Description
params Object
Properties
Name Type Attributes Description
name String The desired short name of the new classifier.
classname_positive_examples ReadStream _positive_examples One or more compressed (.zip) files of images that depict the visual subject for a class within the new classifier. Must contain a minimum of 10 images. You may supply multiple files with different class names in the key.
negative_examples ReadStream <optional>
 A compressed (.zip) file of images that do not depict the visual subject of any of the classes of the new classifier. Must contain a minimum of 10 images. Required if only one positive set is provided.
Source:
Returns:
Type
ReadableStream | undefined

deleteClassifier(params, callback) → {ReadableStream|undefined}

Deletes a custom classifier with the specified classifier id.
Parameters:
Name Type Description
params Object
Properties
Name Type Description
classifier_id String The classifier id
callback function
Source:
Returns:
Type
ReadableStream | undefined

detectFaces(callback) → {ReadableStream|undefined}

Accepts either a url, a single image file, or a zip file with multiple images (.jpeg, .png, .gif) and attempts to extract faces and identities. It then applies a threshold and returns the list of relevant identities, locations, and metadata for found faces for each image. Example output: { "images": [ { "faces": [ { "age": { "max": 54, "min": 45, "score": 0.40459 }, "face_location": { "height": 131, "left": 80, "top": 68, "width": 123 }, "gender": { "gender": "MALE", "score": 0.993307 }, "identity": { "name": "Barack Obama", "score": 0.970688, "type_hierarchy": "/people/politicians/democrats/barack obama" } } ], "image": "obama.jpg" } ], "images_processed": 1 }
Parameters:
Name Type Attributes Description
params.images_file ReadStream <optional>
 The image file (.jpg, .png, .gif) or compressed (.zip) file of images to classify. The total number of images is limited to 100. Either images_file or url must be specified.
params.url String <optional>
 The URL of an image (.jpg, .png, .gif). Redirects are followed, so you can use shortened URLs. The resolved URL is returned in the response. Either images_file or url must be specified.
callback function
Source:
Returns:
Type
ReadableStream | undefined

getClassifier(params, callback) → {ReadableStream|undefined}

Retrieves information about a specific classifier. Example output: { classifier_id: 'fruit_679357912', name: 'fruit', owner: 'a3a48ea7-492b-448b-87d7-9dade8bde5a9', status: 'ready', created: '2016-05-23T21:50:41.680Z', classes: [ { class: 'banana' }, { class: 'apple' } ] }
Parameters:
Name Type Description
params Object
Properties
Name Type Description
classifier_id Boolean The classifier id
callback function
Source:
Returns:
Type
ReadableStream | undefined

listClassifiers(params, callback) → {ReadableStream|undefined}

Retrieve a list of all classifiers, including built-in and user-created classifiers. Example output: {"classifiers": [ { "classifier_id": "fruit_679357912", "name": "fruit", "status": "ready" }, { "classifier_id": "Dogs_2017013066", "name": "Dogs", "status": "ready" } ]}
Parameters:
Name Type Description
params Object
Properties
Name Type Attributes Default Description
verbose Boolean <optional>
 false
callback function
Source:
Returns:
Type
ReadableStream | undefined

recognizeText(callback) → {ReadableStream|undefined}

Accepts either a url, single image file, or a zip file with multiple images (.jpeg, .png, .gif) and attempts to recognize text found in the image. It then applies a threshold and returns the list of relevant locations, strings, and metadata for discovered text in each image. Example output: { "images": [ { "image": "car.png", "text": "3 jag [brio]", "words": [ { "line_number": 0, "location": { "height": 53, "left": 204, "top": 294, "width": 27 }, "score": 0.50612, "word": "3" }, { "line_number": 0, "location": { "height": 32, "left": 264, "top": 288, "width": 56 }, "score": 0.958628, "word": "jag" }, { "line_number": 0, "location": { "height": 40, "left": 324, "top": 288, "width": 92 }, "score": 0.00165806, "word": "brio" } ] } ], "images_processed": 1 }
Parameters:
Name Type Attributes Description
params.images_file ReadStream <optional>
 The image file (.jpg, .png, .gif) or compressed (.zip) file of images to classify. The total number of images is limited to 100. Either images_file or url must be specified.
params.url String <optional>
 The URL of an image (.jpg, .png, .gif). Redirects are followed, so you can use shortened URLs. The resolved URL is returned in the response. Either images_file or url must be specified.
callback function
Source:
Returns:
Type
ReadableStream | undefined