@misc{KoehlerKahraBreuss, author = {K{\"o}hler, Alexander and Kahra, Marvin and Breuß, Michael}, title = {A First Approach to Quantum Logical Shape Classification Framework}, series = {Preprints.org}, journal = {Preprints.org}, publisher = {MDPI AG}, doi = {10.20944/preprints202402.1042.v1}, abstract = {Quantum logic is a well-structured theory, which has recently received some attention because of its fundamental relation to quantum computing. However, the complex foundation of quantum logic borrowing concepts from different branches of mathematics as well as its peculiar settings have made it a non-trivial task to device suitable applications. This article aims to propose for the first time an approach to use quantum logic in image processing at hand of a process for shape classification. We show how to make use of the principal component analysis to realize quantum logical propositions. In this way we are able to assign a concrete meaning to the rather abstract quantum logical concepts, and we are able to compute a probability measure from the principal components. For shape classification we consider encrypting given point clouds of different objects by making use of specific distance histograms. This enables to initiate the principal component analysis. At hand of experiments, we explore the possibility to distinguish between different geometrical objects and discuss the results in terms of quantum logical interpretation.}, language = {en} } @misc{KhanMohammadiSchneidereitMansouriYarahmadietal., author = {Khan Mohammadi, Mohsen and Schneidereit, Toni and Mansouri Yarahmadi, Ashkan and Breuß, Michael}, title = {Investigating training datasets of real and synthetic images for swimmer localisation with YOLO}, series = {Preprints.org}, journal = {Preprints.org}, publisher = {MDPI AG}, doi = {10.20944/preprints202402.0446.v1}, abstract = {In this paper we develop and explore a methodical pipeline for swimmer localisation in outdoor environments. The developed framework is intended to be used for enhancing swimmer safety. A main issue we deal with by the proposed approach is the lack of real world training data in such outdoor environments. Natural lighting changes, dynamic water textures and possibly barely visible swimming persons are key elements to approach. We account for these difficulties by adopting an effective background removal technique with available training data. This allows us to edit swimmers into natural environment backgrounds for the use in subsequent image augmentation. We created 17 training datasets with real images, synthetic images and a mixture of both to investigate different aspects and characteristics of the proposed approach. The datasets are used to train a YOLO architecture for the possible future application in real-time detection. The trained framework is then tested and evaluated on outdoor environment imagery acquired by a safety drone to investigate and confirm the usefulness for outdoor swimmer localisation.}, language = {en} } @misc{ShabaniBreuss, author = {Shabani, Shima and Breuß, Michael}, title = {Semi-Monotone Goldstein Line Search Strategy with Application in Sparse Recovery}, series = {arXiv}, journal = {arXiv}, doi = {https://doi.org/10.48550/arXiv.2503.13099}, pages = {1 -- 12}, abstract = {Line search methods are a prominent class of iterative methods to solve unconstrained minimization problems. These methods produce new iterates utilizing a suitable step size after determining proper directions for minimization. In this paper we propose a semi-monotone line search technique based on the Goldstein quotient for dealing with convex non-smooth optimization problems. The method allows to employ large step sizes away from the optimum thus improving the efficacy compared to standard Goldstein approach. For the presented line search method, we prove global convergence to a stationary point and local R-linear convergence rate in strongly convex cases. We report on some experiments in compressed sensing. By comparison with several state-of-the-art algorithms in the field, we demonstrate the competitive performance of the proposed approach and specifically its high efficiency.}, language = {en} } @misc{ShabaniKhoshghiaferezaeeBreuss, author = {Shabani, Shima and Khoshghiaferezaee, Mohammadsadegh and Breuß, Michael}, title = {Sparse dictionary learning for image recovery by iterative shrinkage}, series = {arXiv}, journal = {arXiv}, doi = {10.48550/arXiv.2503.10732}, pages = {1 -- 19}, abstract = {In this paper we study the sparse coding problem in the context of sparse dictionary learning for image recovery. To this end, we consider and compare several state-of-the-art sparse optimization methods constructed using the shrinkage operation. As the mathematical setting of these methods, we consider an online approach as algorithmical basis together with the basis pursuit denoising problem that arises by the convex optimization approach to the dictionary learning problem. By a dedicated construction of datasets and corresponding dictionaries, we study the effect of enlarging the underlying learning database on reconstruction quality making use of several error measures. Our study illuminates that the choice of the optimization method may be practically important in the context of availability of training data. In the context of different settings for training data as may be considered part of our study, we illuminate the computational efficiency of the assessed optimization methods.}, language = {en} } @misc{SchneidereitGohrenzBreuss, author = {Schneidereit, Toni and Gohrenz, Stefan and Breuß, Michael}, title = {Object detection characteristics in a learning factory environment using YOLOv8}, volume = {2503.10356}, doi = {https://doi.org/10.48550/arXiv.2503.10356}, pages = {1 -- 18}, abstract = {AI-based object detection, and efforts to explain and investigate their characteristics, is a topic of high interest. The impact of, e.g., complex background structures with similar appearances as the objects of interest, on the detection accuracy and, beforehand, the necessary dataset composition are topics of ongoing research. In this paper, we present a systematic investigation of background influences and different features of the object to be detected. The latter includes various materials and surfaces, partially transparent and with shiny reflections in the context of an Industry 4.0 learning factory. Different YOLOv8 models have been trained for each of the materials on different sized datasets, where the appearance was the only changing parameter. In the end, similar characteristics tend to show different behaviours and sometimes unexpected results. While some background components tend to be detected, others with the same features are not part of the detection. Additionally, some more precise conclusions can be drawn from the results. Therefore, we contribute a challenging dataset with detailed investigations on 92 trained YOLO models, addressing some issues on the detection accuracy and possible overfitting.}, language = {en} } @misc{KahraBreuss, author = {Kahra, Marvin and Breuß, Michael}, title = {Colour morphological distance ordering based on the Log-Exp-Supremum}, series = {arXiv}, journal = {arXiv}, doi = {https://doi.org/10.48550/arXiv.2503.11329}, pages = {1 -- 13}, abstract = {Mathematical morphology, a field within image processing, includes various filters that either highlight, modify, or eliminate certain information in images based on an application's needs. Key operations in these filters are dilation and erosion, which determine the supremum or infimum for each pixel with respect to an order of the tonal values over a subset of the image surrounding the pixel. This subset is formed by a structuring element at the specified pixel, which weighs the tonal values. Unlike grey-scale morphology, where tonal order is clearly defined, colour morphology lacks a definitive total order. As no method fully meets all desired properties for colour, because of this difficulty, some limitations are always present. This paper shows how to combine the theory of the log-exp-supremum of colour matrices that employs the Loewner semi-order with a well-known colour distance approach in the form of a pre-ordering. The log-exp-supremum will therefore serve as the reference colour for determining the colour distance. To the resulting pre-ordering with respect to these distance values, we add a lexicographic cascade to ensure a total order and a unique result. The objective of this approach is to identify the original colour within the structuring element that most closely resembles a supremum, which fulfils a number of desired properties. Consequently, this approach avoids the false-colour problem. The behaviour of the introduced operators is illustrated by application examples of dilation and closing for synthetic and natural images.}, language = {en} } @incollection{KahraBreuss, author = {Kahra, Marvin and Breuß, Michael}, title = {Colour morphological distance ordering based on the log-exp-supremum}, series = {Scale space and variational methods in computer vision : 10th international conference, SSVM 2025, Dartington, UK, May 18-22, 2025, proceedings, part II}, booktitle = {Scale space and variational methods in computer vision : 10th international conference, SSVM 2025, Dartington, UK, May 18-22, 2025, proceedings, part II}, publisher = {Springer Nature Switzerland}, address = {Cham}, isbn = {9783031923685}, issn = {0302-9743}, doi = {https://doi.org/10.1007/978-3-031-92369-2_20}, pages = {258 -- 270}, abstract = {Mathematical morphology, a field within image processing, incorporates a variety of filters that either highlight, modify, or eliminate specific information within a mask that traverses the image. Key operations in these filters are dilation and erosion, which determine the supremum or infimum for each pixel with respect to an order of the tonal values over a subset of the image surrounding the pixel. This subset is formed by a structuring element at the specified pixel, which weighs the tonal values. Unlike grey-scale morphology, where tonal order is clearly defined, colour morphology lacks a definitive total order. As no method fully meets all desired properties for colour, because of this difficulty, some limitations are always present. This paper shows how to combine the theory of the log-exp-supremum of colour matrices that employs the Loewner semi-order with a well-known colour distance approach in the form of a pre-ordering. The log-exp-supremum will therefore serve as the reference colour for determining the colour distance. To the resulting pre-ordering with respect to these distance values, we add a lexicographic cascade to ensure a total order and a unique result. The objective of this approach is to identify the original colour within the structuring element that most closely resembles a supremum, which fulfils a number of desired properties. Consequently, this approach avoids the false-colour problem. The behaviour of the introduced operators is illustrated by application examples of dilation and closing for synthetic and natural images.}, language = {en} } @incollection{ShabaniBreuss, author = {Shabani, Shima and Breuß, Michael}, title = {Semi-monotone Goldstein line search strategy with application in sparse recovery}, series = {Scale space and variational methods in computer vision : 10th international conference, SSVM 2025, Dartington, UK, May 18-22, 2025, proceedings, part II}, booktitle = {Scale space and variational methods in computer vision : 10th international conference, SSVM 2025, Dartington, UK, May 18-22, 2025, proceedings, part II}, publisher = {Springer Nature Switzerland}, address = {Cham}, isbn = {9783031923685}, issn = {0302-9743}, doi = {https://doi.org/10.1007/978-3-031-92369-2_6}, pages = {69 -- 81}, abstract = {Line search methods are a prominent class of iterative methods to solve unconstrained minimization problems. These methods produce new iterates utilizing a suitable step size after determining proper directions for minimization. In this paper we propose a semi-monotone line search technique based on the Goldstein quotient for dealing with convex non-smooth optimization problems. The method allows to employ large step sizes away from the optimum thus improving the efficacy compared to standard Goldstein approach. For the presented line search method, we prove global convergence to a stationary point and local R-linear convergence rate in strongly convex cases. We report on some experiments in compressed sensing. By comparison with several state-of-the-art algorithms in the field, we demonstrate the competitive performance of the proposed approach and specifically its high efficiency.}, language = {en} } @misc{CastilloCunninghamWingeretal., author = {Castillo, Susana and Cunningham, Douglas William and Winger, Christian and Breuß, Michael}, title = {Morphological Amoeba-based Patches for Exemplar-Based Inpainting}, series = {Journal of WSCG}, volume = {26}, journal = {Journal of WSCG}, number = {2}, issn = {1213-6972}, pages = {112 -- 121}, language = {en} } @inproceedings{BreussMansouriYarahmadiCunningham, author = {Breuß, Michael and Mansouri Yarahmadi, Ashkan and Cunningham, Douglas William}, title = {The Convex-Concave Ambiguity in Perspective Shape from Shading}, series = {Proceedings of the OAGM Workshop 2018 Medical Image Analysis, May 15 - 16, 2018, Hall/Tyrol, Austria}, booktitle = {Proceedings of the OAGM Workshop 2018 Medical Image Analysis, May 15 - 16, 2018, Hall/Tyrol, Austria}, editor = {Welk, Martin and Urschler, Martin and Roth, Peter M.}, publisher = {Verlag der TU Graz}, address = {Graz}, isbn = {978-3-85125-603-1}, doi = {10.3217/978-3-85125-603-1-13}, pages = {57 -- 63}, abstract = {Shape from Shading (SFS) is a classic problem in computer vision. In recent years many perspective SFS models have been studied that yield useful SFS approaches when a photographed object is close to the camera. However, while the ambiguities inherent to the classical, orthographic SFS models are well-understood, there has been no discussion of possible ambiguities in perspective SFS models. In this paper we deal with the latter issue. Therefore we adopt a typical perspective SFS setting. We show how to transform the corresponding image irradiance equation into the format of the classical orthographic setting by employing spherical coordinates. In the latter setting we construct a convex-concave ambiguity for perspective SFS. It is to our knowledge the first time in the literature that this type of ambiguity is constructed and verified for a perspective SFS model.}, language = {en} } @inproceedings{BreussCunninghamWelk, author = {Breuß, Michael and Cunningham, Douglas William and Welk, Martin}, title = {Scale spaces for cognitive systems: a position paper}, series = {Informatik 2015, Tagung vom 28. September - 2. Oktober 2015 in Cottbus}, booktitle = {Informatik 2015, Tagung vom 28. September - 2. Oktober 2015 in Cottbus}, editor = {Cunningham, Douglas William and Hofstedt, Petra and Meer, Klaus and Schmitt, Ingo}, publisher = {Gesellschaft f{\"u}r Informatik}, address = {Bonn}, isbn = {978-3-88579-640-4}, pages = {1253 -- 1255}, language = {en} } @misc{HajighasemiBreuss, author = {Hajighasemi, Saeide and Breuß, Michael}, title = {Fuzzy Frankot-Chellappa algorithm for surface normal integration}, series = {Algorithms}, volume = {18}, journal = {Algorithms}, number = {8}, publisher = {MDPI AG}, address = {Basel}, issn = {1999-4893}, doi = {10.3390/a18080488}, pages = {1 -- 19}, abstract = {In this paper, we propose a fuzzy formulation of the classic Frankot-Chellappa algorithm by which surfaces can be reconstructed using normal vectors. In the fuzzy formulation, the surface normal vectors may be uncertain or ambiguous, yielding a fuzzy Poisson partial differential equation that requires appropriate definitions of fuzzy derivatives. The solution of the resulting fuzzy model is approached by adopting a fuzzy variant of the discrete sine transform, which results in a fast and robust algorithm for surface reconstruction. An adaptive defuzzification strategy is also introduced to improve noise handling in highly uncertain regions. In experiments, we demonstrate that our fuzzy Frankot-Chellappa algorithm achieves accuracy on par with the classic approach for smooth surfaces and offers improved robustness in the presence of noisy normal data. We also show that it can naturally handle missing data (such as gaps) in the normal field by filling them using neighboring information.}, language = {en} } @misc{ShabaniBreussKahraetal., author = {Shabani, Shima and Breuß, Michael and Kahra, Marvin and Teiser, Jens and V{\"o}lke, Gretha Swantje and Wenders, Nico}, title = {Morphological granulometric analysis of particle imagery from microgravity experiments}, series = {arXiv}, journal = {arXiv}, publisher = {Cornell University}, address = {Ithaca, NY}, doi = {10.48550/arXiv.2508.06593}, pages = {1 -- 12}, abstract = {The aim of our work is to analyze size distributions of particles and their agglomerates in imagery from astrophysical microgravity experiments. The data acquired in these experiments are given by sequences consisting of several hundred images. It is desirable to establish an automated routine that helps to assess size distributions of important image structures and their dynamics in a statistical way. The main technique we adopt to this end is the morphological granulometry. After preprocessing steps that facilitate granulometric analysis, we show how to extract useful information on size of particle agglomerates as well as underlying dynamics. At hand of the discussion of two different microgravity key experiments we demonstrate that the granulometric analysis enables to assess important experimental aspects. We conjecture that our developments are a useful basis for the quantitative assessment of microgravity particle experiments.}, language = {en} } @misc{KoehlerBreuss, author = {K{\"o}hler, Alexander and Breuß, Michael}, title = {Recognition of geometrical shapes by dictionary learning}, series = {Proceedings of International Conference on Artificial Intelligence, Computer, Data Sciences and Applications (ACDSA 2025), 7-9 August 2025, Antalya-T{\"u}rkiye}, volume = {2025}, journal = {Proceedings of International Conference on Artificial Intelligence, Computer, Data Sciences and Applications (ACDSA 2025), 7-9 August 2025, Antalya-T{\"u}rkiye}, publisher = {IEEE}, address = {Piscataway, New Jersey}, isbn = {979-8-3315-3562-9}, doi = {10.1109/ACDSA65407.2025.11166282}, pages = {1 -- 6}, abstract = {Dictionary learning is a versatile method to produce an overcomplete set of vectors, called atoms, to represent a given input with only a few atoms. In the literature, it has been used primarily for tasks that explore its powerful representation capabilities, such as for image reconstruction. In this work, we present a first approach to make dictionary learning work for shape recognition, considering specifically geometrical shapes. As we demonstrate, the choice of the underlying optimization method has a significant impact on recognition quality. Experimental results confirm that dictionary learning may be an interesting method for shape recognition tasks.}, language = {en} } @misc{HajighasemiBreuss, author = {Hajighasemi, Saeide and Breuß, Michael}, title = {An improved fuzzified Frankot and Chellappa method for surface normal integration}, series = {2025 IEEE International Conference on Fuzzy Systems (FUZZ) : conference proceedings, July 06-09 2025, Reims, France}, journal = {2025 IEEE International Conference on Fuzzy Systems (FUZZ) : conference proceedings, July 06-09 2025, Reims, France}, publisher = {IEEE}, address = {Piscataway, NJ}, isbn = {979-8-3315-4319-8}, doi = {https://doi.org/10.1109/FUZZ62266.2025.11197608}, pages = {1 -- 6}, abstract = {We present an enhanced fuzzified variant of the classical Frankot-Chellappa algorithm for reconstructing surfaces from normal vector fields by using a multiplicative fuzzification process and an adaptive defuzzification strategy. By incorporating local spread metrics and noise-level estimation, the adaptive defuzzification process dynamically adjusts reconstruction weights, significantly improving robustness to noise. Experimental results demonstrate that the proposed method achieves superior accuracy compared to a previous fuzzy approach, particularly under noisy conditions.}, language = {en} } @incollection{ShabaniBreussKahraetal., author = {Shabani, Shima and Breuß, Michael and Kahra, Marvin and Teiser, Jens and V{\"o}lke, Gretha Swantje and Wenders, Nico}, title = {Morphological granulometric analysis of particle imagery from microgravity experiments}, series = {Discrete geometry and mathematical morphology : 4th International Joint Conference, DGMM 2025, Groningen, The Netherlands, November 3-6, 2025, proceedings}, booktitle = {Discrete geometry and mathematical morphology : 4th International Joint Conference, DGMM 2025, Groningen, The Netherlands, November 3-6, 2025, proceedings}, publisher = {Springer Nature Switzerland}, address = {Cham}, isbn = {978-3-032-09544-2}, issn = {0302-9743}, doi = {https://doi.org/10.1007/978-3-032-09544-2_35}, pages = {487 -- 500}, abstract = {The aim of our work is to analyze size distributions of particles and their agglomerates in imagery from astrophysical microgravity experiments. The data acquired in these experiments are given by sequences consisting of several hundred images. It is desirable to establish an automated routine that helps to assess size distributions of important image structures and their dynamics in a statistical way. The main technique we adopt to this end is the morphological granulometry. After preprocessing steps that facilitate granulometric analysis, we show how to extract useful information on size of particle agglomerates as well as underlying dynamics. At hand of the discussion of two different microgravity key experiments we demonstrate that the granulometric analysis enables to assess important experimental aspects. We conjecture that our developments are a useful basis for the quantitative assessment of microgravity particle experiments.}, language = {en} } @misc{ShabaniKhoshghiaferezaeeBreuss, author = {Shabani, Shima and Khoshghiaferezaee, Mohammadsadegh and Breuß, Michael}, title = {Sparse dictionary learning for image recovery by iterative shrinkage}, series = {Intelligent systems and applications : proceedings of the 2025 Intelligent Systems Conference (IntelliSys) : volume 2}, journal = {Intelligent systems and applications : proceedings of the 2025 Intelligent Systems Conference (IntelliSys) : volume 2}, editor = {Arai, Kohei}, publisher = {Springer Nature Switzerland}, address = {Cham}, isbn = {978-3-032-00071-2}, issn = {2367-3370}, doi = {https://doi.org/10.1007/978-3-032-00071-2_19}, pages = {309 -- 327}, abstract = {In this research we study the sparse coding problem in the context of sparse dictionary learning for sparse image recovery. To this end, we consider and compare several state-of-the-art sparse nonsmooth optimization methods constructed using the shrinkage operation. As the mathematical setting of these methods, we consider an online approach as algorithmical basis together with the basis pursuit denoising problem that arises by the convex optimization approach to the dictionary learning problem. By a dedicated construction of datasets and corresponding dictionaries, we study the effect of enlarging the underlying learning database on reconstruction quality making use of several error measures. Our study illuminates that the choice of the optimization method may be practically important in the context of availability of training data. In the context of different settings for training data as may be considered part of our study, we illuminate the computational efficiency of the assessed optimization methods.}, language = {en} } @misc{SchneidereitGohrenzBreuss, author = {Schneidereit, Toni and Gohrenz, Stefan and Breuß, Michael}, title = {Object detection characteristics in a learning factory environment using YOLOv8}, series = {Intelligent systems and applications : proceedings of the 2025 Intelligent Systems Conference (IntelliSys) : volume 2}, volume = {2}, journal = {Intelligent systems and applications : proceedings of the 2025 Intelligent Systems Conference (IntelliSys) : volume 2}, publisher = {Springer Nature Switzerland}, address = {Cham}, isbn = {978-3-032-00071-2}, issn = {2367-3370}, doi = {https://doi.org/10.1007/978-3-032-00071-2_18}, pages = {288 -- 308}, abstract = {AI-based object detection, and efforts to explain and investigate their characteristics, is a topic of high interest. The impact of, e.g., complex background structures with similar appearances as the objects of interest, on the detection accuracy and, beforehand, the necessary dataset composition are topics of ongoing research. In this study, we present a systematic investigation of background influences and different features of the object to be detected. The latter includes various materials and surfaces, partially transparent and with shiny reflections in the context of an Industry 4.0 learning factory. Different YOLOv8 models have been trained for each of the materials on different sized datasets, where the appearance was the only changing parameter. In the end, similar characteristics tend to show different behaviours and sometimes unexpected results. While some background components tend to be detected, others with the same features are not part of the detection. Additionally, some more precise conclusions can be drawn from the results. Therefore, we contribute a challenging dataset with detailed investigations on 92 trained YOLO models, addressing some issues on the detection accuracy and possible overfitting.}, language = {en} } @misc{KhoshghiaferezaeeKrauthShabanietal., author = {Khoshghiaferezaee, Mohammadsadegh and Krauth, Moritz Frederic and Shabani, Shima and Breuß, Michael}, title = {Quality versus sparsity in image recovery by dictionary learning using iterative shrinkage}, series = {2025 Fourteenth International Conference on Image Processing, Theory, Tools \& Applications (IPTA)}, journal = {2025 Fourteenth International Conference on Image Processing, Theory, Tools \& Applications (IPTA)}, publisher = {IEEE}, address = {Piscataway, NJ}, isbn = {978-1-6654-5739-2}, issn = {2154-512X}, doi = {10.1109/IPTA66025.2025.11222043}, pages = {1 -- 6}, abstract = {Sparse dictionary learning (SDL) is a fundamental technique that is useful for many image processing tasks. As an example we consider here image recovery, where SDL can be cast as a nonsmooth optimization problem. For this kind of problems, iterative shrinkage methods represent a powerful class of algorithms that are subject of ongoing research. Sparsity is an important property of the learned solutions, as exactly the sparsity enables efficient further processing or storage. The sparsity implies that a recovered image is determined as a combination of a number of dictionary elements that is as low as possible. Therefore, the question arises, to which degree sparsity should be enforced in SDL in order to not compromise recovery quality. In this paper we focus on the sparsity of solutions that can be obtained using a variety of classic and modern iterative shrinkage optimization algorithms. It turns out that there are three different sparsity regimes that occur depending on the method in use. Furthermore, we illustrate that high sparsity does in general not compromise recovery quality, even if the recovered image is quite different from the learning database. In addition, we present tests that confirm the generalization properties of our findings.}, language = {en} } @misc{KahraBreussKleefeldetal., author = {Kahra, Marvin and Breuß, Michael and Kleefeld, Andreas and Welk, Martin}, title = {Matrix-valued LogSumExp approximation for colour morphology}, series = {Journal of mathematical imaging and vision}, volume = {67}, journal = {Journal of mathematical imaging and vision}, number = {5}, publisher = {Springer US}, address = {New York}, issn = {0924-9907}, doi = {10.1007/s10851-025-01267-5}, abstract = {Mathematical morphology is a part of image processing that employs a moving window to modify pixel values through the application of specific operations. The supremum and infimum are pivotal concepts, yet defining them in a general sense for high-dimensional data such as colour is a challenging endeavour. As a result, a number of different approaches have been taken to try to find a solution, with certain compromises being made along the way. In this paper, we present an analysis of a novel approach that replaces the supremum within a morphological operation with the LogExp approximation of the maximum for matrix-valued colours. This approach has the advantage of extending the associativity of dilation from the one-dimensional to the higher-dimensional case. Furthermore, the minimality property is investigated and a relaxation specified to ensure that the approach is continuously dependent on the input data.}, language = {en} } @misc{EltaherBreuss, author = {Eltaher, Mahmoud and Breuß, Michael}, title = {Deep learning for unsupervised 3D shape representation with superquadrics}, series = {AI}, volume = {6}, journal = {AI}, number = {12}, publisher = {MDPI AG}, address = {Basel}, issn = {2673-2688}, doi = {10.3390/ai6120317}, pages = {1 -- 36}, abstract = {The representation of 3D shapes from point clouds remains a fundamental challenge in computer vision. A common approach decomposes 3D objects into interpretable geometric primitives, enabling compact, structured, and efficient representations. Building upon prior frameworks, this study introduces an enhanced unsupervised deep learning approach for 3D shape representation using superquadrics. The proposed framework fits a set of superquadric primitives to 3D objects through a fully integrated, differentiable pipeline that enables efficient optimization and parameter learning, directly extracting geometric structure from 3D point clouds without requiring ground-truth segmentation labels. This work introduces three key advancements that substantially improve representation quality, interpretability, and evaluation rigor: (1) A uniform sampling strategy that enhances training stability compared with random sampling used in earlier models; (2) An overlapping loss that penalizes intersections between primitives, reducing redundancy and improving reconstruction coherence; and (3) A novel evaluation framework comprising Primitive Accuracy, Structural Accuracy, and Overlapping Percentage metrics. This new metric design transitions from point-based to structure-aware assessment, enabling fairer and more interpretable comparison across primitive-based models. Comprehensive evaluations on benchmark 3D shape datasets demonstrate that the proposed modifications yield coherent, compact, and semantically consistent shape representations, establishing a robust foundation for interpretable and quantitative evaluation in primitive-based 3D reconstruction.}, language = {en} }