Himafor ISTTS again held an online webinar on Saturday, September 5, 2020, to be precise at 1:00 p.m. with the topic "Implementing Computer Vision is SUPER Easy with Google MLKit". Internal participants participated in this webinar through the Zoom platform. This webinar was hosted by Malvin Patrick Kurniawan, a 2016 class of ISTTS alumnus who now works as a Full Stack Engineer at PT Djarum. The resource person opens the webinar with an introduction first, then continues with an explanation of Google MLKit.
For mobile developers, it might be very easy with the Google MLKit. Google MLKit is an SDK (library) that Google created for Android and iOS applications. There are many advantages that can be obtained from MLKit such as no need for an internet connection, some of its features can be built-in by Android or iOS and are also easy to use so we don't have to make modeling or training first. There are 3 major features that have been supported, namely the Vision API. This feature will have a lot to do in the vision area, such as side detection, face detection, barcode scanning. The second feature is Natural Language API. This feature deals with text, such as language identification. The last feature is the Custom Model. In this feature, we can create our own modeling by sending the model that we have created to Google MLKit
There are many features provided by Google MLKit. One of the highlights is Face Detection. Machine Learning related to Face Detection must be able to provide any features on our faces such as eyes, ears, nose, or mouth. MLKit must also be able to provide facial contours (points on our faces). For example, we want to detect the mouth, then the mouth will be marked with dots on the mouth. The specialty of this feature is that Google MLKit will provide information about the expression on a person's face. Is the person smiling or the left eye is closed. The great thing is because this feature does not need an internet connection and can also detect faces in real time. Another advantage that is obtained from this feature is that it can detect many faces at once.
At the end of the material, the resource person presents a snippet of the program in Android Studio. "Google MLKit can be a solution to create a cool but easy computer vision application without having a CS (Computer Science) background by using dots process image to get the results," said the source. The webinar then ended with a question and answer session between the participants and the resource persons. Then proceed with a group photo and fill in attendance for participants who have attended the webinar.