Hello Fernando,
Please try the following:
import PhotoScan
doc = PhotoScan.app.document
chunk = doc.activeChunk
p1 = PhotoScan.app.getInt("Specify first camera index:", 0)
p2 = PhotoScan.app.getInt("Specify second camera index:", len(chunk.cameras)-1)
p1 = chunk.cameras[p1]
p2 = chunk.cameras[p2]
region = PhotoScan.Region()
dist = p2.center - p1.center
region.size = PhotoScan.Vector([dist.x, dist.y, dist.z])
region.center = (p1.center + p2.center) / 2.
region.rot = PhotoScan.Matrix().diag([1,1,1])
chunk.region = region
PhotoScan.app.update()
On script run you'll be asked about camera indices you wish to use (leading zero).