Google’s AI/Machine Learning Tool Demonstrates How Algorithm Understand An Image
Google published a new tool to analyze your images, which tells you what it thinks about the image is relevant for, by using machine learning/artificial intelligence algorithm. This tool is called Google Cloud Vision Image Analysis Tool, indicates AI and Machine Learning algorithms for understanding images.
What is Google Cloud Vision Image Analysis Tool?
Google’s Cloud Vision API is a cloud service that allows you to add image analysis features to apps and websites. The tool offers two computer vision products that use Google’s AI and Machine Learning algorithms, help you understand your images with industry-leading prediction accuracy.
This tool use advanced machine learning models, allows you to upload an image, and quickly classify them into millions of pre-defined categories. This new Google’s Cloud Vision Image Analysis Tool can detect objects, faces, read printed and handwritten text, and build valuable metadata into your image catalog.
Google’s Image Analysis Tools Classifications:
- Web Entities
- Safe Search
The “faces” tab detects the facial attributes such as emotional state or wearing head-wear, but facial Recognition is not supported. As you can see when I uploaded my picture, which is clearly smiling but Google’s image analysis tool didn’t detect it correctly.
Yes, the “objects” tab works perfectly and detect what objects are using in the image, like ears, nose, mouth, but also detects like portrait and photography.
The “labels” tab extract information about entities within an image that Google recognizes, like jacket, material, smile, hairs and etc. It works very well.
This ‘Web’ tab detects web references to an image, which means it shows how much images are related on the web. Google’s Cloud Vision Image Analysis Tool uses multiple criteria for understanding an image.
The Image Properties tab detects general attributes of the image, such as dominant color.
Safe Search tab detects how the image explicit content such as adult content or violent content within an image. This feature uses five categories (“adult”, “spoof”, “medical”, “violence”, and “racy”), and returns the likelihood that each is present in a given image.
This is an amazing Image Analysis Tool for getting an idea that how Google understand your images. It may also helpful for you to optimize the images in the better ways.
Want to know what Google thinks about your image? Find out here at https://cloud.google.com/vision/docs/drag-and-drop