🤖 AI Summary
To address the challenge of high-precision, single-image-driven 3D reconstruction for minimally invasive brain surgery—where conventional point cloud methods suffer from sparsity and low accuracy, and densely annotated data are scarce—this paper proposes a two-stage stereo-aware graph-structured GAN. The first stage generates coarse-grained organ topology; the second stage employs a parameter-free attention-guided free-form transformation module to upsample region-of-interest details and correct geometric defects. We introduce the first end-to-end framework integrating graph neural networks with adversarial training, optimized geometrically via Chamfer distance. Our method achieves significant improvements over state-of-the-art approaches: it reduces PC-to-PC error and Chamfer distance, increases reconstructed point cloud density by 3.2×, enhances visual fidelity, and boosts downstream classification accuracy by 4.7%.
📝 Abstract
In minimally-invasive brain surgeries with indirect and narrow operating environments, 3D brain reconstruction is crucial. However, as requirements of accuracy for some new minimally-invasive surgeries (such as brain-computer interface surgery) are higher and higher, the outputs of conventional 3D reconstruction, such as point cloud (PC), are facing the challenges that sample points are too sparse and the precision is insufficient. On the other hand, there is a scarcity of high-density point cloud datasets, which makes it challenging to train models for direct reconstruction of high-density brain point clouds. In this work, a novel model named stereoscopic-aware graph generative adversarial network (SG-GAN) with two stages is proposed to generate fine high-density PC conditioned on a single image. The Stage-I GAN sketches the primitive shape and basic structure of the organ based on the given image, yielding Stage-I point clouds. The Stage-II GAN takes the results from Stage-I and generates high-density point clouds with detailed features. The Stage-II GAN is capable of correcting defects and restoring the detailed features of the region of interest (ROI) through the up-sampling process. Furthermore, a parameter-free-attention-based free-transforming module is developed to learn the efficient features of input, while upholding a promising performance. Comparing with the existing methods, the SG-GAN model shows superior performance in terms of visual quality, objective measurements, and performance in classification, as demonstrated by comprehensive results measured by several evaluation metrics including PC-to-PC error and Chamfer distance.
