This depends on the type of memory you are talking about.
In general though, the idea of pixels and 'saving' an image like a computer is not very accurate. Our eyes don't have an even distribution of photoreceptors, only a very small portion of our view is in focus and there are big obstructions (a blind spot where the nerve endings leave the eyeball, obscured vision by the nose, etc.). What we perceive is not "pixels" being picked up by the eye but an image that is reconstructed by the brain. Our perception is our brain's interpretation rather than raw data coming in (that is why it is subject to illusions, like the checker shadow illusion).
We have a sensory memory store that helps us actually see a stable image and is also responsible for things like after images or seeing a swinging sparkler as a circle of light (again our perception is an integration and interpretation rather than the raw data of the sparkler moving). Some of that can even happen unconsciously.
When it comes to remembering an image we saw a while ago, there is evidence that we reconstruct it in a backwards way, starting wit the main gist first. It is more like placing the important objects there and then filling in the details with what we remember. Interestingly, every time we recall a memory the reconstruction influences our subsequent recall, which can result in completely false memories. Say, we remember seeing a red car on a rainy day. When we remember this image, we immediately remember the red car but may not have much recollection of the weather. The brain may remember the red very lusciously because it stood out to us rather than because it was bright and falsely interpret that as it being a bright and sunny day. With each recall, we may fill in the weather more towards the sunny side until we are fully convinced it was a sunny day.
On the whole, visual memory is not like pixels but more like object reconstruction. Hope this helps
References:
Wikipedia (n.d.) Checker Shadow Illusion https://en.wikipedia.org/wiki/Checker_shadow_illusion
Pang, D. K., & Elntib, S. (2021). Strongly masked content retained in memory made accessible through repetition. Scientific reports, 11(1), 1-10. https://doi.org/10.1038/s41598-021-89512-w
Linde-Domingo, J., Treder, M. S., Kerrén, C., & Wimber, M. (2019). Evidence that neural information flow is reversed between object perception and object reconstruction from memory. Nature communications, 10(1), 1-13. https://doi.org/10.1038/s41467-018-08080-2
Neuroscience News (2020). Brain’s ‘updating mechanisms’ may create false memories https://neurosciencenews.com/updating-mechanism-false-memory-16438/