Line segment detection is a crucial building block in visual SLAM, 3D reconstruction, and industrial inspection. Recent deep learning methods have significantly improved accuracy, yet even the smallest models require several megabytes of memory, exceeding the capacity of low-cost MCUs. This work investigates the maximum achievable accuracy under a sub-megabyte budget and proposes MiLSD, a detector tailored for MCU-level constraints.
We systematically compare three output representations within a compact fully-convolutional backbone. Our study shows that the proposed F-Clip center-with-length-and-angle formulation learns most effectively at small model sizes. We find that 8-bit quantization preserves full-precision performance, while 4-bit quantization causes significant degradation, particularly in angle regression, with quantization-aware training recovering only part of the loss.
With a one-megabyte activation budget and inference enhancements including sub-pixel decoding, test-time augmentation, and a lightweight verifier, MiLSD improves sAP10 on ShanghaiTech Wireframe from 10.6 (25k parameters, 0.25 MB) to 24.1 within 1 MB. Rather than competing with GPU-scale parsers, we map the accuracy-memory trade-off across representations, bit-widths, capacities, and post-processing strategies for embedded vision systems.