Superpixel Integrated Grids for Fast Image Segmentation
This addresses computational bottlenecks in image segmentation for researchers and practitioners, offering a more efficient alternative without sacrificing accuracy, though it is incremental as it builds on existing superpixel methods.
The paper tackles the problem of inefficient deep learning for image segmentation with superpixels by introducing SIGRID, a data structure that encodes superpixel color and shape to reduce input dimensionality; results show it matches or surpasses pixel-level performance on benchmark datasets while accelerating training.
Superpixels have long been used in image simplification to enable more efficient data processing and storage. However, despite their computational potential, their irregular spatial distribution has often forced deep learning approaches to rely on specialized training algorithms and architectures, undermining the original motivation for superpixelations. In this work, we introduce a new superpixel-based data structure, SIGRID (Superpixel-Integrated Grid), as an alternative to full-resolution images in segmentation tasks. By leveraging classical shape descriptors, SIGRID encodes both color and shape information of superpixels while substantially reducing input dimensionality. We evaluate SIGRIDs on four benchmark datasets using two popular convolutional segmentation architectures. Our results show that, despite compressing the original data, SIGRIDs not only match but in some cases surpass the performance of pixel-level representations, all while significantly accelerating model training. This demonstrates that SIGRIDs achieve a favorable balance between accuracy and computational efficiency.