googleCloudVisionR: Access to the 'Google Cloud Vision' API for Image Recognition, OCR and Labeling

Description

Interact with the 'Google Cloud Vision' <https://cloud.google.com/vision/> API in R. Part of the 'cloudyr' <https://cloudyr.github.io/> project.

Author(s)

Maintainer: Jeno Pal paljenczy@gmail.com

Authors:

• Tamas Koncz t.koncz@gmail.com

• Balazs Varkoly varkoly.balazs@gmail.com

• Peter Lukacs lukacs.peter.andras@gmail.com

• Eszter Kocsis p.kocsis.eszter@gmail.com

Other contributors:

• Florian Teschner [contributor]

See Also

Useful links:

• Report bugs at https://github.com/cloudyr/googleCloudVisionR/issues

helper function to send POST request to the Google Vision API

Description

sends the request defined in 'body' to the API

Usage

call_vision_api(body, apiEndpoint = "images:annotate",
  httpRequestType = "POST")

Arguments

Value

API response in raw format

helper function to create json for response request

Description

creates a json output from the inputs

Usage

create_request_body(imagePaths, feature, maxNumResults)

Arguments

Value

request body (payload), encoded as json

helper function to create a list of details of one image annotation request

Description

creates a list output from the inputs

Usage

create_single_image_request(imagePath, feature, maxNumResults)

Arguments

Value

list of request details for one image

helper function to base64 encode the image file

Description

base64 encodes an image file

Usage

encode_image(imagePath)

Arguments

Value

get the image back as encoded file

helper function code to extract the annotations

Description

a utility to extract features from the API response

Usage

extract_annotations(responses, imagePaths, featureType)

Arguments

Value

a data.table

helper function code to extract error from API response into a data.table

Description

helper function code to extract error from API response into a data.table

Usage

extract_error(responses, imagePaths)

Arguments

Value

a data.table

helper function code to extract the response data.frame

Description

a utility to extract features from the API response

Usage

extract_response(responses, imagePaths, feature)

Arguments

Value

a data.table

helper function code to provide an extractor function for different feature types

Description

a utility to provide functions to extract features from the API response

Usage

extractor(featureType)

Arguments

Value

a function

helper function code to extract API response into a data.table for given feature type

Description

helper function code to extract API response into a data.table for given feature type

Usage

face_detection_extractor(response)

Arguments

Value

a data.table

helper function code to record available feature types

Description

helper function code to record available feature types

Usage

gcv_get_available_feature_types()

Value

a list of available features names and their types (as returned by the API)

Examples

gcv_get_available_feature_types()

Get parsed image annotations from the Google Cloud Vision API

Description

Given a list of images, a feature type and the maximum number of responses, this functions calls the Google Cloud Vision API, and returns the image annotations in a data.table format.

Usage

gcv_get_image_annotations(imagePaths, feature = "LABEL_DETECTION",
  maxNumResults = NULL, batchSize = 64L, savePath = NULL)

Arguments

Value

a data frame with image annotation results

Examples

## Not run: 
    # Label Detection (default), with maximum 7 results returned per image
    imagePath <- system.file(
      "extdata", "golden_retriever_puppies.jpg", package = "googleCloudVisionR"
    )
    gcv_get_image_annotations(imagePaths = imagePath, maxNumResults = 7)

    # Face detection
    imagePath <- system.file(
      "extdata", "arnold_wife.jpg", package = "googleCloudVisionR"
    )
    gcv_get_image_annotations(imagePaths = imagePath, feature = "FACE_DETECTION")

    # Google Cloud Storage URI as input
    gcv_get_image_annotations("gs://vision-api-handwriting-ocr-bucket/handwriting_image.png")

## End(Not run)

Get raw API response from the Google Cloud Vision API

Description

Given a list of images, a feature type and the maximum number of responses, this functions calls the Google Cloud Vision API, and returns the raw response from the API. For a friendlier response, refer to the 'gcv_get_image_annotations' function, which returns results in a data.table format (however, the information returned is limited compared to the raw response).

Usage

gcv_get_raw_response(imagePaths, feature = "LABEL_DETECTION",
  maxNumResults = NULL)

Arguments

Value

a response object returned by the API. To get the image annotations, take the "content" element from the object

Examples

## Not run: 
    imagePath <- system.file(
      "extdata", "golden_retriever_puppies.jpg", package = "googleCloudVisionR"
    )
    raw_response <- gcv_get_raw_response(imagePaths = imagePath, maxNumResults = 7)

    str(raw_response)
    raw_response[["content"]]

## End(Not run)

helper function to call the API for one batch of images

Description

helper function to call the API for one batch of images

Usage

gcv_get_response(imagePaths, feature, maxNumResults)

Arguments

Value

a data frame with image annotation results

helper function code to extract Bounding Box x,y coordinates for an API response element

Description

helper function code to extract Bounding Box x,y coordinates for an API response element

Usage

get_bounding_boxes(response)

Arguments

Value

a data.table

helper function to validate input image paths

Description

helper function to validate input image paths

Usage

get_invalid_image_paths(vec)

Arguments

Value

vector of invalid paths from @vec

helper function code to extract API response into a data.table for given feature type

Description

helper function code to extract API response into a data.table for given feature type

Usage

label_detection_extractor(response)

Arguments

Value

a data.table

helper function code to extract API response into a data.table for given feature type

Description

helper function code to extract API response into a data.table for given feature type

Usage

landmark_detection_extractor(response)

Arguments

Value

a data.table

helper function code to extract API response into a data.table for given feature type

Description

helper function code to extract API response into a data.table for given feature type

Usage

logo_detection_extractor(response)

Arguments

Value

a data.table

helper function code to extract API response into a data.table for given feature type

Description

helper function code to extract API response into a data.table for given feature type

Usage

ocr_extractor(response)

Arguments

Value

a data.table

helper function to split a vector to approximately equally sized chunks

Description

helper function to split a vector to approximately equally sized chunks

Usage

split_to_chunks(vec, chunkSize)

Arguments

Value

a list of chunks