> I need to get image data through USB port and refresh the panel with frame buffer 480*(800/8)=48000 bytes.
I expected something like that.
I mean, an application expected to shuffle around large data volumes at high(er) throughput.
Just keep in mind that generic MCUs are not very well suited for that.
DSPs are made for that, and most application processors (Cortex A) and some high-end Cortex M devices have dedicated hardware to handle displays.
With a generic MCU, you need to accept limitations.
You could reduce the package size, and optimize it through double-buffering mechanisms and/or DMA usage.
I would suggest to profile your application, and see where it spends most time.
And evaluate what options you have to improve performance, or reduce the visible impact.