Keep your eyes in the cross between the two images.
This gif is demonstrating that peripheral vision is really creepy. Our pattern recognition that’s hardwired to recognize faces puts together the components or groups of components of a face whether upside down or right side up.
The brain is selectively juxtaposing certain aspects of the individual features on the left with those on the right because you’re crossing your eyes forcing your brain to have to reassemble what it’s seeing. But your facial recognition is kicking in so it’s combining specific facial components not just whole faces or shapes, contours, etc.
Peripheral vision is a part of vision that occurs outside the very center of gaze. There is a broad set of non-central points in the field of view that is included in the notion of peripheral vision. “Far peripheral” vision exists at the edges of the field of view, “mid-peripheral” vision exists in the middle of the field of view, and “near-peripheral”, sometimes referred to as “para-central” vision, exists adjacent to the center of gaze.