5 Main Types of Image Annotation

hello and welcome to machine learning 101 today we'll be talking about computer vision and five types of image annotation but we've only got five minutes so let's get started in the world of ai and machine learning data is king without data there can be no data science and one of the largest demands for training data comes from a sector of artificial intelligence called computer vision put simply computer vision is the field of ai that seeks to give machines the ability to see the world and understand visual data the same way humans do how do we do that well we start with a set of algorithms that are designed to analyze and process image data these sets of algorithms are called neural networks and there are many different kinds of neural networks that are suited to different tasks you could have a neural network that is designed to recognize human faces or one that is good at tracking movements in videos or even some that can detect cancer in x-rays while there are many different kinds of neural networks one thing that almost all of them have in common is an endless need for training data humans learn new concepts by going to school studying reading books or even watching videos and neural networks need to study too self-driving cars for example need to be able to detect traffic lights pedestrians and other vehicles [Music] but we can't just give the neural network a bunch of images and expect it to learn on its own just like how humans need a teacher certain neural networks need to be taught what to look for in the images and that's where image annotation comes in image annotation is simply the process of attaching labels to an image this can arrange from one label for the entire image or numerous labels for groups of pixels within the image and there are five different types of image annotation commonly used in computer vision today number one is bounding box annotation for bounding boxes we simply draw a box around certain objects within the image the box should be drawn as close as possible to the outer edges of the object number two is 3d bounding box or cuboid annotation much like bounding boxes with cuboid annotation we draw a box around objects in an image however cuboids can show length width and approximate depth of the target objects number three is polygon annotation it is often the case that numerous objects in an image weren't designed to fit perfectly within a box with polygons annotators draw lines by placing dots around the outer edge of the object that they want to annotate the process is like a connect the dots exercise while placing the dots at the same time the space within the area surrounded by the dots is then annotated using a predetermined set of classes next we have lines and splines while lines and splines can be used for a variety of purposes they are mainly used to train machines to recognize lanes and boundaries as their name suggests annotators would simply draw lines along the boundaries you require the neural network to learn lines and splines can be used to train warehouse robots to accurately place boxes in a row or items on a conveyor belt however the most common application of lines and splines annotation is autonomous vehicles by annotating road lanes and sidewalks the autonomous vehicle can be trained to understand batteries and stay in one lane without veering number five is semantic segmentation whereas the previous examples on this list dealt with outlining the outer edges of an object semantic segmentation is much more precise it is the process of associating every single pixel in an entire image with a label to do this we use a process similar to polygon annotation where we draw lines around the group of pixels we want to label each segment is usually indicated by a unique color code the end result would look something like this semantic segmentation can also be used for incredibly specialized tasks like tagging brain lesions within ct scan images so these were just five common types of image annotation used in computer vision today hopefully this tutorial helped you have a better grasp of the basics behind computer vision and image annotation before we go please note that this video was brought to you by lionbridge ai if you're looking to build your own image data sets get in touch to find out how lionbridge can assist you with a community of one million data scientists and professional annotators lionbridge can create custom data sets to meet your project's needs thank you for watching machine learning 101 if you liked what you saw please help us out by liking the video and for more machine learning tutorials under 5 minutes please hit the subscribe button below that's all see you next time

Channel: Machine Learning 101

Views: 18,479

Keywords: image annotation, computer vision, machine learning, ai, data science

Id: YJ5HpEILiDE

Length: 5min 1sec (301 seconds)

Published: Wed Aug 26 2020