Against Expectations: A Simple Greedy Heuristic Outperforms Advanced Methods in Bitmap Decomposition

Na minha lista:
Detalhes bibliográficos
Publicado no:Electronics vol. 14, no. 13 (2025), p. 2615-2653
Autor principal: Pitkäkangas Ville
Publicado em:
MDPI AG
Assuntos:
Acesso em linha:Citation/Abstract
Full Text + Graphics
Full Text - PDF
Tags: Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!

MARC

LEADER 00000nab a2200000uu 4500
001 3229143380
003 UK-CbPIL
022 |a 2079-9292 
024 7 |a 10.3390/electronics14132615  |2 doi 
035 |a 3229143380 
045 2 |b d20250101  |b d20251231 
084 |a 231458  |2 nlm 
100 1 |a Pitkäkangas Ville 
245 1 |a Against Expectations: A Simple Greedy Heuristic Outperforms Advanced Methods in Bitmap Decomposition 
260 |b MDPI AG  |c 2025 
513 |a Journal Article 
520 3 |a Partitioning rectangular and rectilinear shapes in n-dimensional binary images into the smallest set of axis-aligned n-cuboids is a fundamental problem in image analysis, pattern recognition, and computational geometry, with applications in object detection, shape simplification, and data compression. This paper introduces and evaluates four deterministic decomposition methods: pure greedy selection, greedy with backtracking, greedy with a priority queue, and an iterative integer linear programming (IILP) approach. These methods are benchmarked against three established baseline techniques across 13 diverse 1D–4D images (up to 8 × 8 × 8 × 8 elements), featuring holes, concavities, and varying orientations. Surprisingly, the simplest approach—a purely greedy heuristic selecting the largest unvisited region at each step—consistently achieved optimal or near-optimal decompositions, even for complex images, and maintained optimality under rotation without post-processing. By contrast, the more sophisticated methods (backtracking, prioritization, and IILP) exhibited trade-offs between speed and quality, with IILP adding overhead without superior results. Runtime testing showed IILP was on average ~37× slower than the fastest greedy method (ranging from ~3× to 100× slower). These findings highlight that a well-designed greedy strategy can outperform more complex algorithms for practical binary shape decomposition, offering a compelling balance between computational efficiency and solution quality in pattern recognition and image analysis. 
653 |a Heuristic 
653 |a Linear programming 
653 |a Image analysis 
653 |a Integer programming 
653 |a Pattern analysis 
653 |a Data compression 
653 |a Optimization 
653 |a Decomposition 
653 |a Methods 
653 |a Algorithms 
653 |a Image quality 
653 |a Object recognition 
653 |a Pattern recognition 
653 |a Computational geometry 
773 0 |t Electronics  |g vol. 14, no. 13 (2025), p. 2615-2653 
786 0 |d ProQuest  |t Advanced Technologies & Aerospace Index 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3229143380/abstract/embedded/ITVB7CEANHELVZIZ?source=fedsrch 
856 4 0 |3 Full Text + Graphics  |u https://www.proquest.com/docview/3229143380/fulltextwithgraphics/embedded/ITVB7CEANHELVZIZ?source=fedsrch 
856 4 0 |3 Full Text - PDF  |u https://www.proquest.com/docview/3229143380/fulltextPDF/embedded/ITVB7CEANHELVZIZ?source=fedsrch