Image-focusing analysis can provide useful velocity information,
in particular in areas where conventional velocity
analysis lacks resolution.
Measuring image coherency across the structural-dip axes
provides quantitative information on image focusing.
However, in the presence of curved reflectors
coherency across dips suffers similar shortcomings
as conventional minimum entropy functional.
To overcome these difficulties,
I propose to perform image-focusing analysis
by measuring image coherency simultaneously across
structural dips and reflection aperture/azimuth angles.
The application of the proposed image-focusing semblance
to two synthetic data sets
demonstrates its potential as a tool to extract
quantitative velocity information from image-focusing analysis.