It has been around two decades, In 2001 when Google launched Google Images to search for images over the world wide web. A decade after that Google introduced reverse image search in 2011 which help you to search using an image rather than a keyword. And now Google have launched it’s cloud vision tool. Though a lot of features where available to be use from before only. The main ones are:
- Vision API:-Google vision tool will allow us to analyze features of an image using it’s AI & Machine Learning algorithms which can be used in our apps and websites. This has majorly 7 categories like Faces, Objects, Labels etc which we will be learning about in this article only
- AutoML Vision:- So, Your kind of data doesn’t fit in the type created by Vision API or you are a kind of person who wants to keep things in your own hands. Here comes Google’s AutoML vision into play. You can train and create your own Machine Learning models according to your labels and let the model learn from it and then it can be used to further classify images based on the labels provided by you.
- Vision Product Search:- Own a Shopify store? Are you a merchant? Then you might have got you might be searching for. According to Google, “Cloud Vision Product Search allows retailers to create a set of products, each containing reference images that visually describe the product from a set of viewpoints.” Currently, Vision API Product Search supports the following product categories: Homegoods, apparel, and toys.
Now, Let’s explore what Vision API has for us. There are mainly 10 ways provided in which Vision API can be utilized.
- Faces:– Using this you can identify the number of faces in a particular picture. You can also get internals like where is the retina, chin etc as the response from the API. Point to be noted that Facial Recognition is not supported. It can also detect facial attributes such as emotional state or wearing headwear. Even Facebook use these kinds of models in which can be observed in the images alt text.
- Logos:- You can also identify logos of major companies and you would be able to get the data about it. In response, you will get vertices of the logo and name of the company to which the logo belongs to. This can also work for multiple logos.
- Objects:- Well, It is quite simple in this case. You can identify objects along with the probability of its occurrence.
- Labels: These are basically like tags which are relevant to the images. It can be any aspects related to it.
- Web: This basically shows descriptive words which are associated with the image over the web. In short, All web references to the image.
- Image Properties: This basically tells about the colours used in the image and crop pointers.
- Safe Search: This can be very useful if you want to identify an image is adult or not. Especially in case of a forum or community where it is quite hectic to take care of items being posted by users. It can be of great help.
- Text: This basically gives you all the text it can read from a picture. Almost like Google Goggles.
- Document Text: Vision API is also giving you a way by which it can read text from an image. Also, It performs Optical Character Recognition which can also read document including handwriting.
- Landmarks: With this, we can detect popular natural and man-made structures within an image.
With all these Google is giving you a lot of power. It is your time to create some cool stuff and show us. Go ahead and test it on a browser-based tool here. If we need help with implementing the API, Do let us know in comments, we will provide a guide for doing so in future. The trial for the API is free through the API itself is paid. If you are up to creating something serious that won’t hurt much. To know more about its evolution go here.
Do let us know what you are about to do next with Google vision API. Till then, Keep Uppskilling because #RuknaTohHaiHiNahi