Automate business processes with visual search engines and cloud-based storage

This article will look at how an organization automates business processes using visual search engines and cloud-based storage.

Mobile applications that use visual search engines are becoming increasingly important. As technology matures, more and more use cases are being developed in the areas of defence, insurance, medical care and fashion. The ability to photograph and use an algorithm to identify objects in an image requires a data store for the algorithm to perform comparisons, and the data stores are gradually transferred to the cloud. This article will look at the available visual search engine algorithms, learn how they use data storage, connect your applications to these data stores, and choose the pros and cons of a specific vendor solution.

What is a visual search engine?

Although visual search engines are not limited to mobile devices (smartphones or tablets), they are the most common terminal or user interface, as mobile devices now have built-in cameras. With these cameras, mobile applications can interact asynchronously with the images taken by the camera.

By using a visual search engine, a user can take a two-dimensional picture and use the search algorithm to determine whether an image contains recognizable objects. These algorithms are deployed to mobile applications through software connectors called application programming interfaces (APIs). The APIs from visual search engine vendors such as IQ engines enable programmers to create their own applications using visual search engine technology.

Some visual search engine vendors have some pre-built mobile applications available, such as Google Goggles. But Google also needs to deploy an API for Goggles, which means there are restricted use cases for deploying applications to the industry. An Italian company called Macroglossa also offers visual search engines. The deployment or use of Macroglossa is little known in the industry, but it is an alternative to IQ engines and Google. Regardless of the visual search engines used, they are related to the general dynamic capabilities and processes.

Using the IQ engines Visioniq as an example, a user can take a picture using a mobile device, which triggers a client-side visual search process when the application is loaded with Visioniq. The visual search engine API service then invokes the server-side software, referencing the business rules generated by the cloud-based data store. Finally, for public data storage, a crowdsourcing (crowdsourcing) approach can be enabled, allowing the public to help refine the search algorithm. Figure 1 illustrates the process.

Figure 1. IQ engines Visioniq defined visual search engine flow

For visual search engines used to identify images, users must first train (training) the system with a combination of images and supporting attributes, and then upload them to a cloud-based data store or to a web crawler (crawler). Google Goggles used the web crawler route. To use Goggles, you must first convert the image and support attributes or metadata to HTML. Visioniq leverages cloud-based data storage, where developers and technicians upload content to the data store through the representational state transmits (REST)-ful API. Regardless of the image and how the metadata associated with the image (color, texture, date, and brand) is scanned or uploaded, the algorithm must analyze and continue to refine the data. Figure 2 provides an image sample with related properties.

Figure 2. Visioniq Data Model

Note that search algorithms and properties must be refined repeatedly in the process of allocating progress that allows for appropriate success rates. After training the system to an acceptable rate of accuracy, you can start to deploy the user acceptance test (UAT). UAT enables developers and organizations to improve image and search algorithms, making applications more useful in real-world scenarios such as light, blur screens, and multiple angles. Because the cloud is often used to store data that trains these systems, it's wise to figure out what a cloud is.

Combining cloud computing with visual search
Now, let me explain how to use cloud computing and visual search engines together.