Not reliably from the image alone. You can estimate it, but you usually cannot determine a single definitive “true resolution.”

A common approach is frequency analysis, such as a Fourier transform: if an image has been upscaled, the highest spatial frequencies may be weak or missing. But that only shows the file lacks fine detail — it does not prove why. The same result can come from soft focus, motion blur, noise reduction, diffraction, low-detail subject matter, or compression.

Metadata can sometimes help. EXIF may identify the camera model and expected native image sizes, which can reveal whether the file dimensions are unusual. Image forensics tools may also detect signs of resizing, recompression, cloning, or editing from JPEG tables or noise patterns.

Still, without trustworthy metadata or an original file, any answer is statistical rather than certain. In short: software can suggest whether detail is limited or whether resizing likely occurred, but it generally cannot state an exact “actual resolution” with confidence from a processed photo alone.