Technology that automatically corrects wide-angle lens distortion

Wide-angle camera lenses, which are also increasing on smartphones, may prevent the person at the edge of the image from being captured. This is because the wide-angle lens itself becomes a shape that stretches the edges and edges on the road. However, a research team developed a solution that automatically corrects only these distortions.

A wide-angle lens looks like a slightly protruding center and can capture a wider field of view than a typical smartphone lens with a relatively flat shape. Therefore, if you are taking a selfie, a wide-angle lens can capture a person in a wide range. However, the farther from the center of the picture, the more the image is stretched, and the selfie viewed with a wide-angle lens becomes distorted as if it were drooping to anyone reflected in the corners and edges.

Of course, editing tools such as Photoshop can correct these distortions, but it is not software that anyone can handle without burden. Because the average smartphone user is not an experienced Photoshop user. However, researchers at MIT CSAIL (Computer Science & Artificial Intelligence Lab) and Google have developed an algorithm that automatically corrects facial distortions at these corners and edges without any other partial distortion.

To correct edge distortion, you need to create a grid corresponding to the content content and detect whether the recognized face is distorted. Make sure that the distortion correction does not affect other parts, then adjust the distortion correction so that it blends well with the surroundings. For example, the straight line of the building in the background does not bend due to the crystal.

Algorithm calculations and modifications were fully automated. Thanks to this, users can finish editing with a single button in the editing app. Moreover, by integrating into the camera app itself, photos taken with a wide-angle lens can be instantly modified. As such, the algorithm calculation seems to be light.

It is not yet known when this feature will actually be implemented on smartphones. However, all researchers who participated in the study are currently working at Google. If it is installed on a smartphone, there is a high possibility that the pixel phone will be number one. Related information can be found here .