FPGA and MCU are not interchangeable marketing words. They are engineering tools at different levels. The FPGA suits real-time, high-speed, parallel and timing-sensitive work: display data organisation, pixel mapping, interface bridging, row/column timing, receiving-side output and test-pattern generation.
The MCU suits state management, peripheral control, communication protocol, parameter storage, fault records, programming and production test. In a real project the two usually work together: the FPGA owns the real-time display path, the MCU owns configuration, monitoring and maintenance.
The judgement that matters is whether a project needs an FPGA at all, whether an MCU alone is enough, whether an existing main controller plus an interface board would do, or whether a full sending/receiving chain is required. Small cost-sensitive products should not be over-designed; high-resolution or multi-module projects should not be forced through a low-cost control path.