Speaker: Majid Banaeyan (PRIP)
As panoramic images are widely used in many applications, efficient image stitching methods that provide visually pleasant image mosaics are needed. We are working on a novel concept for smart camera image stitching based on graph pyramids. For a multi-camera system, the images have to be aligned accordingly to create an image mosaic. Instead of calculating the corresponding transformations centrally, we aim at enabling each camera to individually calculate the transformation of the image it takes. Graph pyramids used for image segmentation provide information about the segmentation process. We analyze how this information can be used to calculate the transformations for image alignment.