No audio! Examples of Filters usage for meaningful objects detection.
No audio!
From my personal feelings - humans detect objects on a visual scene trying to apply a pattern on a 3-d model (or 2-D if this is literally a picture).
This applies all non-visual understanding of the surrounding: where we are at, what is expected surrounding, which objects can be there and so on. For example, if I'm in the forest, I expect a specific color palette and specific set of objects (trees, bushes, anymals, birds, grass, e.t.c) around me. This is why people are so easily lost in "foreign" forests.
For computer vision we should first segment visual scenes to start to apply patterns to objects, and people trying to do a "universal" solution that will not have human-like background understanding for the surrounding.
Too sad. Too complex.
For example: a photo has different brightness zones (not visible by an eye) on different parts of the picture, so segmentation is complicated by the fact that the same pixel color is part of objects (plates) and of boundaries (dirt + grass).
Solution is:
To apply filters to filter out objects or boundaries. Sequence filters are an option!
BTW, just a thought: until AI has this "non visual surrounding understanding", like "I see a forest, where expected objects are ..1...2...3" - don't call it an Intelligence...