Dedicated camera image processors are typically application-specific chips designed in hardware for a narrow set of tasks such as image processing, JPEG encoding, and video compression. In that sense they are more like specialized accelerators than a general-purpose CPU.

A CPU is built to handle a very wide range of unknown tasks, so it trades specialization for flexibility. A dedicated image processor includes hardware optimizations or instructions for the exact operations a camera repeatedly performs. Because of that, for the same power, heat, and clock-speed limits, a dedicated image processor will generally outperform a general-purpose CPU at those specific jobs.

A mobile processor like a Snapdragon may share some architectural similarities and can certainly process images in software, but without the same task-specific hardware it usually won’t be as efficient at the camera-specific workload. A general-purpose processor can only catch up by using much higher clock speeds or more parallel hardware, which increases cost and power use.

So yes: the idea is similar to hardware acceleration, but camera image processors are tuned specifically for imaging tasks rather than broad computing.