FLaTEC: Frequency-Disentangled Latent Triplanes for Efficient Compression of LiDAR Point Clouds
Published in IEEE International Conference on Multimedia and Expo (ICME), 2026
FLaTEC presents a frequency-aware LiDAR point cloud compression framework that projects voxel features into compact triplanes and disentangles low- and high-frequency components for flexible bitrate allocation. With progressive frequency modulation and lightweight local refinement, it preserves fine structural details while enabling real-time compression and achieving state-of-the-art rate-distortion performance on SemanticKITTI and Ford benchmarks.
Recommended citation: Xiaoge Zhang, Zijie Wu, Mingtao Feng, Zichen Geng, Mehwish Nasim, Saeed Anwar, & Ajmal Mian. (2026). FLaTEC: Frequency-Disentangled Latent Triplanes for Efficient Compression of LiDAR Point Clouds. 2026 IEEE International Conference on Multimedia and Expo (ICME).
Download Paper