Stay organized with collections
Save and categorize content based on your preferences.
Learn about Vision changes such as backward incompatible
API changes, product or feature deprecations, mandatory migrations, or
potentially disruptive maintenance.
An improved model is now available for
Text Detection (OCR). The new model
can be used with TEXT_DETECTION and
DOCUMENT_TEXT_DETECTION features. The same model is used for
requests sent to both features.
With the new model, the distribution of
confidence scores of responses will change. More information
below.
Please note that you have 90 days from today to test the new model by
specifying "builtin/latest" in the model field of the
Feature
object. At the end of that period, it will be promoted to the default model
accessible as "builtin/stable". After that event, the original models will
still be available for another 90 days using "builtin/legacy".
If you encounter problems with this upgrade, please contact the Vision API
engineering team by submitting a ticket in the
private issue tracker.
October 01, 2021
Differences in the distribution of confidence scores of responses
("builtin/latest" vs. "builtin/stable").
FNR (False Negative Ratio): The ratio of correct symbol predictions with
confidence score less than the threshold.
FPR (False Positive Ratio): The ratio of incorrect symbol predictions with
confidence score greater than or equal to the threshold.
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Hard to understand","hardToUnderstand","thumb-down"],["Incorrect information or sample code","incorrectInformationOrSampleCode","thumb-down"],["Missing the information/samples I need","missingTheInformationSamplesINeed","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-09-04 UTC."],[],[],null,["# Service announcements\n\nLearn about Vision changes such as backward incompatible\nAPI changes, product or feature deprecations, mandatory migrations, or\npotentially disruptive maintenance.\n\nOctober 01, 2021\n----------------\n\nDifferences in the distribution of confidence scores of responses\n(\"builtin/latest\" vs. \"builtin/stable\").\n\n**FNR (False Negative Ratio)**: The ratio of correct symbol predictions with\nconfidence score less than the threshold.\n\n**FPR (False Positive Ratio)**: The ratio of incorrect symbol predictions with\nconfidence score greater than or equal to the threshold."]]