• Edge and primitive feature extraction serve as core primitives for early vision systems: local operators detect edges and corners, gradient-based edge cues are organized into lines/curves, and transform-based grouping (e.g., Hough) enables robust scene interpretation. This theme aggregates edge-focused work that underpins higher-level pattern analysis [1], [4], [11], [15].

• Texture features and statistical pattern recognition drive image classification and scene understanding, with textural descriptors and probabilistic feature spaces shaping early content-based and pattern-recognition pipelines [2], [3], [10].

• 3D surface inference from 2D imagery and stereo frames combines multi-view cues, disparity information, and shape descriptors to recover surfaces and depth structure from images [8], [9], [13], [14], [18].

• Partial-object recognition emphasizes robustness to occlusion via local-feature emphasis and focused search strategies, enabling locating partially visible objects under challenging views [4], [7], [16].

• Digital image processing and computational imaging create the foundational processing, representation, and compression workflows enabling subsequent vision tasks and feature extraction [5], [6], [19].