Abstract:In the field of referring segmentation, top-down methods and bottom-up methods are the two prevailing approaches. Both of these methods inevitably exhibit certain drawbacks. Top-down methods are susceptible to Polar Negative (PN) errors due to their limited understanding of multi-modal fine-grained features. Bottom-up methods lack macro-level object positional information, making them susceptible to Inferior Positive (IP) errors. However, we find that the two approaches are highly complementary in addressing their respective weaknesses, but combining them directly through a simple average does not yield complementary advantages. Therefore, we proposed a hierarchical collaboration approach to explore the complementary characteristics of the existing two methods from the perspectives of fusion and interaction, aiming to achieve more precise segmentation results. We proposed the Complementary Feature Interaction (CFI) module, which enables top-down methods to access fine-grained information and allows bottom-up approaches to obtain object positional information interactively. Regarding integration, Gaussian Scoring Integration (GSI) models the Gaussian performance distributions of two branches and performs weighted integration by sampling confidence scores from these distributions. We integrate various top-down and bottom-up methods within the proposed architecture and conduct experiments on three standard datasets. The experimental results demonstrate that our method outperforms the state-of-the-art independent segmentation algorithms. On the RefCOCO validation, test A and test B datasets, our proposed method achieved IoU scores of 77.51, 79.12, and 72.79, respectively. Extensive experiments demonstrate that our method can significantly improve segmentation accuracy when fusing different sub-methods.

Contour Detection and Hierarchical Image Segmentation

Decompose Image into Meaningful Regions Based on Contour Detector and Watershed Algorithm

Efficient Local and Global Contour Detection Based on Superpixels.

Contour Based Automatic Scene Segmentation in Image Sequences

Image Segmentation Using Hierarchical Merge Tree

Contour Segment Grouping for Object Detection.

An efficient hierarchical graph based image segmentation

Generalizing Edge Detection to Contour Detection for Image Segmentation

AdaContour: Adaptive Contour Descriptor with Hierarchical Representation

Human Segmentation With Deep Contour-Aware Network

Hierarchical Contour Closure-Based Holistic Salient Object Detection

Contour Detection Based on Binocular Parallax Perception Mechanism

Salient Object Detection and Segmentation

Hierarchical Object-Oriented Video Segmentation and Representation Algorithm

Deep ContourFlow: Advancing Active Contours with Deep Learning

A Hierarchical Segmentation Method for Multispectral Imagery

Hierarchical Image Segmentation Based on Multi-feature Fusion and Graph Cut Optimization.

Color And Region Based Alternate-Hierarchical Image Segmentation Algorithm

A Two-Stage Image Segmentation Via Global and Local Region Active Contours

Hierarchical Collaboration for Referring Image Segmentation

Improved techniques for automatic image segmentation