Google Cloud Vision API

I explored another new service offered by Google Cloud. This is for extracting information from and about images it is called Google Cloud Vision. This service offers powerful features like Optical Character Recognition, Identifying objects in the images, location information etc. You can see from the screenshots shown below what details you can extract from the images using this service. The first tab shows the highest matching category for the images. The image has lot of colors and it also echoes the theme of play and that it can be a banner image. The second tab shows the text within the image. As you can see it uses OCR component to extract text from the image. The third tab shows what are the dominant colors in the image. The next tab shows that it is a safe search image. If you want to automate this process you can do it in any client language and get the response in JSON format as shown in the last tab.

In addition it can also analyze face and emotions and other aspects if the image has human beings or animals in the picture. Before you start using this service I will describe how this service is billed. It is not billed per image but per feature you request for. The following are the features that can be requested for your images.

• Label Detection (Categories for your images)
• Optical Character Recognition (OCR)
• Facial Detection (Identifying person in the image)
• Landmark Detection (Detect the geo location where the picuture is taken)
• Logo detection
• Image Properties like color, safe search etc.

If you request OCR and Facial Detection then you are requesting 2 features for your images and hence you will be billed for 2 uints. Google allows upto 1000 units per month free. After the first 1000 units you will be billed $1.50 per 1000 units. If you are planning to use this service then list the features that you are planning to use and then do a rough calculation before executing the service to scan your images.