Google Translate camera can detect 88 languages

The camera in the Google Translate app is now equipped with an automatic language detection feature.

This feature, as quoted from Engadget, Monday (15/7/2019), is capable of detecting 88 languages ranging from the languages of countries in Europe to Africa and Asia.

The way it works is simple; The user only needs to point the camera to the object containing the text. Then the Google Translate app will automatically detect the language of the writing and display its translation.

For those who often travel abroad, this feature will be very useful. For example, to read the menu when ordering meals in the restaurant or reading street names or directions.

The technology behind this feature is Neural Machine Translation (NMT). The use of NMT for camera in the Google Translate app originated in 2015. While the use of NMT for Google Translate Web version started since 2016.

Neural Machine Translation

Neural Machine Translation is a translation approach that uses a large neural network to predict possible sequences of words, often also in full sentence form.

Compared to Statistical Machine Translation (SMT), which takes up more memory and time, NMT trains its parts to maximize performance. Thus, the NMT can outperform the traditional tunneling approach.

In accuracy, NMT is claimed to reduce the error rate from 55 to 85 percent on certain source language pairs and get specific languages. In addition to Google Translate app and web versions, Google has also been using the NMT on Google Lens.

In comparison, in mid-2018 then LinkedIn introduced the translation feature. The number of languages supported by this professional network reaches 60 languages.

0 Response to "Google Translate camera can detect 88 languages"

Post a Comment

Iklan Atas Artikel

Iklan Tengah Artikel 1

Iklan Tengah Artikel 2

Iklan Bawah Artikel