The article discusses how computers can understand shapes in images. It looks at different types of shape models and how they can be used to recover 3D shapes. The researchers studied fully parametric models and found that using a specific measure can help get more accurate results. They also looked at partially parametric models and developed a method to recover 3D shapes without needing to know certain details. The study shows that by using certain techniques, computers can better understand and describe shapes in the real world.