3D Scanning Research Articles

Primate Phenotypes: A Multi-Institution Collection of 3D Morphological Data Housed in MorphoSource

The field of phenomics is experiencing unprecedented advances thanks to the rapid growth of morphological quantification based on three-dimensional (3D) imaging, online data repositories, team-oriented collaborations, and open data-sharing policies. In line with these progressions, we present an extensive primate phenotypic dataset comprising >6,000 3D scans (media) representing skeletal morphologies of 386 individual specimens covering all hominoid genera (except humans) and other selected primates. The digitized specimens are housed in physical collections at the American Museum of Natural History, the National Museum of Natural History, the Royal Museum for Central Africa (Belgium), the Cleveland Museum of Natural History, and Stony Brook University. Our technical validation indicates that despite the diverse digitizing devices used to produce the scans, the final 3D models (meshes) can be safely combined to collect comparable morphometric data. The entire dataset (and detailed associated metadata) is freely available through MorphoSource. Hence, these data contribute to empowering the future of primate phenomics and providing a roadmap for future digitization and archiving of digital data from other collections.

水稻籽粒堆积特性离散元法模拟研究

Accurately determining the angle of repose for irregular and dispersed agricultural grain materials requires a simulation model that effectively represents the actual grain shapes and utilizes numerical methods to analyze their stacking behavior. This study focuses on "Yongyou 15" rice grains, employing 3D raster scanning technology to obtain precise contour data. Through a reverse modeling process, a detailed 3D geometric model of the grains was developed, resulting in a discrete element model comprising 618 grains of varying diameters, created using granular polymer theory. Discrete element analysis software (EDEM) was integrated with MATLAB image processing to simulate the falling and stacking process of the rice grains within a stainless steel bottomless cylindrical tube. The contour of the grain heap was analyzed using linear fitting, followed by a micro-mechanical investigation of the grain heap structure. The analysis indicated that the pressure depression within the heap is caused by the oblique transmission of contact forces. The simulated angle of repose under experimental conditions was 31.29°±0.41°, differing by only 0.81% from the actual measured angle of 31.04°±0.21° obtained through physical stacking experiments. These results demonstrate that combining numerical simulations with image feature extraction is a reliable and efficient method for assessing the stacking properties of agricultural materials.

Augmented reality for rhinoplasty: 3D scanning and projected AR for intraoperative planning validation

AbstractRhinoplasty is one of the major surgical procedures most popular and it is generally performed modelling the internal bones and cartilage using a closed approach to reduce the damage of soft tissue, whose final shape is determined by means of their new settlement over the internal remodelled rigid structures. An optimal planning, achievable thanks to advanced acquisition of 3D images and thanks to the virtual simulation of the intervention via specific software. Anyway, the final result depends also on factors that cannot be totally predicted regarding the settlement of soft tissues on the rigid structures, and a final objective check would be useful to eventually perform some adjustments before to conclude the intervention. The main idea of the present work is the using of 3D scan to acquire directly in the surgical room the final shape of the nose and to show the surgeon the differences respect to the planning in an intuitive way using augmented reality (AR) to show false colours directly over the patient face. This work motivates the selection of the devices integrated in our system, both from a technical and an ergonomic point of view, whose global error, evaluated on an anthropomorphic phantom, is lower than ± 1.2 mm with a confidence interval of 95%, while the mean error in detecting depth thickness variations is 0.182 mm.

Point-Cloud Instance Segmentation for Spinning Laser Sensors

In this paper, we face the point-cloud segmentation problem for spinning laser sensors from a deep-learning (DL) perspective. Since the sensors natively provide their measurements in a 2D grid, we directly use state-of-the-art models designed for visual information for the segmentation task and then exploit the range information to ensure 3D accuracy. This allows us to effectively address the main challenges of applying DL techniques to point clouds, i.e., lack of structure and increased dimensionality. To the best of our knowledge, this is the first work that faces the 3D segmentation problem from a 2D perspective without explicitly re-projecting 3D point clouds. Moreover, our approach exploits multiple channels available in modern sensors, i.e., range, reflectivity, and ambient illumination. We also introduce a novel data-mining pipeline that enables the annotation of 3D scans without human intervention. Together with this paper, we present a new public dataset with all the data collected for training and evaluating our approach, where point clouds preserve their native sensor structure and where every single measurement contains range, reflectivity, and ambient information, together with its associated 3D point. As experimental results show, our approach achieves state-of-the-art results both in terms of performance and inference time. Additionally, we provide a novel ablation test that analyses the individual and combined contributions of the different channels provided by modern laser sensors.

The use of digital technologies in assessing the technical condition of historic structures

This article explores the use of modern digital technologies, such as 3D scanning and photogrammetry, in assessing the technical condition of historic structures. It compares traditional and digital inventory methods, emphasising the advantages of the latter in the precise and rapid acquisition of spatial data. Three types of laser scanning are described – simplified, handheld, and stationary – along with their applications in monument documentation. Case studies are presented where digital technologies were applied to the analysis and conservation of structures such as the ruins of Melsztyn Castle, Czersk Castle, the historic brickworks in Izbica, and the Juliusz Osterwa Theatre in Lublin. The findings demonstrate that these technologies enhance inventory accuracy, enable the detection of damage invisible to the naked eye, and save time and resources. Challenges related to processing large volumes of data and the need for specialised knowledge and standards are also discussed. In conclusion, the application of digital technologies in cultural heritage protection offers significant benefits and is invaluable for preserving monuments for future generations.

The Use of iPad LiDAR to Build a Digital Elevation Model in Defining Customary Zones (Case Study of Panglipuran, Bali)

Abstract. The LiDAR on the iPad device can be used for surveying and mapping. However, the accuracy and precision of the LiDAR scanner on the iPad are not as high as the LiDAR system on Terrestrial Laser Scanning (TLS), with a more limited range and smaller coverage area. Still, the LiDAR sensor integrated into the iPad has advantages when compared to professional LiDAR systems, including the iPad LiDAR as a low-cost portable 3D scanner, dynamic portability, speed, easy use at close range so that it can be used to scan objects with high density and can reach areas/parts of objects that are difficult to be covered by TLS. The 3D modelling method using iPad LiDAR is a solution for building 3D models.Since iPad’s point cloud data is not integrated with the global coordinate system, a georeferencing process is necessary to produce DEM (Digital Elevation Model), which will be used to define customary zones in Panglipuran Heritage Village. The georeferencing process is carried out by transforming the iPad’s LiDAR coordinates system to the global reference coordinates resulting from GNSS measurements. The scanning strategy must be well arranged. In addition, good control point design determines the quality of control point transformation results into a global reference coordinate system.This research contributes to obtaining an alternative low-cost 3D mapping method for cultural conservation areas using LiDAR technologies. This method is inexpensive but still considers geospatial quality aspects so it can be used for the decision-making process in spatial planning of cultural regions.

3D Non-separable Moment Invariants and Their Use in Neural Networks

Recognition of 3D objects is an important task in many bio-medical and industrial applications. The recognition algorithms should work regardless of a particular orientation of the object in the space. In this paper, we introduce new 3D rotation moment invariants, which are composed of non-separable Appell moments. We show that non-separable moments may outperform the separable ones in terms of recognition power and robustness thanks to a better distribution of their zero surfaces over the image space. We test the numerical properties and discrimination power of the proposed invariants on three real datasets—MRI images of human brain, 3D scans of statues, and confocal microscope images of worms. We show the robustness to resampling errors improved more than twice and the recognition rate increased by 2–10 % comparing to most common descriptors. In the last section, we show how these invariants can be used in state-of-the-art neural networks for image recognition. The proposed H-NeXtA architecture improved the recognition rate by 2–5 % over the current networks.

Using Digital Methods to Work on an Existing Sculptural Object or One in The Process of Physical Fabrication

The report examines key aspects of using digital methods in working on sculptural objects, both existing and in progress. An analysis of current trends and the impact of technological advances in sculpture is presented, with a focus on 3D scanning and virtual modelling Digital methods and tools and software platforms and hardware technologies for scanning, modelling and virtual creation are explored. Through a review of real-world examples and research projects, the possibilities and advantages of digital technologies in sculpture are illustrated. It analyses the challenges that arise when using digital methods and highlights the importance of combining classical sculptural techniques and digital skills. Emphasising the importance of digital methods as tools that enrich and extend sculptural practice and create opportunities for achieving high quality and innovation in art.

The integrate geomorphological, morphotectonics investigations using remote sensing data from space as the basis for the efficiency increasing of geological works

The underestimation of the integrated application of geomorphological, morphotectonic information and materials of space remote data during a geological research, the geological survey and predictive-prospecting work significantly reduces their effectiveness. The imperfection of the geomorphological researches methodology focused on the study of territories relief as a collection of surfaces of geological bodies is one of the possible reasons for this. Perspectives opens the transition to the research of landforms and geological structures, bodies in their unity, as three-dimensional, volumetric objects. Fundamentally new opportunities for studying the geology of the seas and oceans bottom are associated, in particular, with the visualization of digital 3D relief models created on the basis of the global bathymetric database GEBCO 2014 (http ://Ocean3dproects...), regional echo-depth-sounder metering’s maps, altimetry materials and remote sensing of the Earth from space. The topicality of such investigations is also due to the fact determined that plate tectonics cannot explain many features of the structure of the Earth and individual regions relief. A number of examples show that modern geomorphology combined with methods of remote sensing from space and computerization should be considered as one of the necessary disciplines for geological mapping and work aimed at solving various geological problems. The further common development of the planet relief sciences (geomorphology + morphotectonics) determines the need to train geomorphologists at both geographical and geological faculties of universities.

Development of a Methodology for Obtaining Solid Models of Products That Are Objects of Reverse Engineering Using the Example of the Capstone Micro-GTU C 65

Currently, about a thousand micro gas turbine units of small and medium capacity are in operation in the Russian Federation, which are used as an autonomous power source at critical infrastructure facilities. During long-term operation, the component parts of the micro GTU may fail and require replacement or repair. The lack of spare parts and design documentation for their production makes it impossible to operate. As a way to solve the problem, the reverse engineering process can be used to produce components. One of the stages of reverse engineering is to determine the geometric parameters of the object. The fastest and most accurate way to obtain geometric characteristics in the reverse engineering process is 3D scanning. Three-dimensional scanning technology is used to obtain a solid 3D model of the prototype surface, based on which design documentation is subsequently developed. This article presents the results of a study of the influence of the parameters of the distance between polygonal grid points and the scanner exposure on the detailing of the outer surface and the geometric parameters of the resulting polygonal model. As a result of this study, the dependence of the final file size and the time spent on scanning and processing on the distance between the points of the polygonal grid and the model was established. Based on the dependence of the parameters, recommendations were obtained for choosing the distance between the points of the polygonal grid of laser 3D scanning. Also, after performing the stages of reverse engineering, the methodology for creating solid models and design documentation of parts of power equipment units using 3D scanning technology was improved.

Dynamically programable real-time controller for a 2D scanning fibre microscope

A programmable controller for a 2D Lissajous scanning fibre microscope is described. Source motion is based on a vibrating cantilever formed by dip-coating two cylindrical silica fibres. Orthogonal modes are excited by a piezoelectric actuator oriented at 45° to the principal axes. Back-scattered signals are detected using a mode-stripping photodiode to collect cladding modes in a dual numerical aperture confocal scheme. Optical feedback is generated by a shaped reflecting aperture with amplitude-coded reflectivity. Electrical pulses from the detector are separated into low- and high-mode feedback signals using window detectors. The low mode is excited at resonance using a phase-locked loop (PLL) containing a voltage-controlled oscillator (VCO). The high mode is driven off-resonance at a frequency derived from computer control of a second VCO. Amplitudes are corrected, and common frequency signals derived from the two motions using divide-by-N circuits are synchronised using additional PLLs. Programmable generation of Lissajous figures and imaging with dynamically variable scan density are demonstrated.

Diagnostic value of digital breast tomosynthesis in suspicious lesions detected in screening mammogram

BackgroundDigital mammography (DM) serves as the primary breast imaging technique for early cancer detection and diagnosis. However, it faces certain limitations, particularly in imaging dense breasts. DM exhibits reduced sensitivity and specificity in women with radiographically dense breast tissue due to diminished contrast between potential tumors and surrounding tissue, as well as tissue summation potentially obscuring lesions. The aim of this study was to assess the diagnostic efficacy of digital breast tomosynthesis compared to digital mammography in evaluating suspicious breast lesions. The study included 60 female patients eligible for mammography. All participants underwent full-field digital mammography and 3d tomosynthesis. Written informed consent was obtained from all subjects after explaining the research’s purpose and benefits. The Ethics Committee of the Faculty of Medicine, Menoufia University approved the study under code no. 10/2022RADI24.ResultsRegarding BIRADS classification: All 60 patients were categorized using the BIRADS lexicon for both mammography and digital breast tomosynthesis. Mammography results showed BIRADS 3 in 10% (6/60), BIRADS 4a in 33.3% (20/60), 4b in 10% (6/60), BIRADS 4c in 26.7% (16/60), and BIRADS 5 in 20% (12/60) of cases. In comparison, tomosynthesis classified cases as BIRADS 3 in 13.3% (8/60), 4b in 16.7% (10/60), BIRADS 4c in 23.3% (14/60), and BIRADS 5 in 46.7% (28/60) of cases.ConclusionsDigital breast tomosynthesis (DBT) has established itself as a significant advancement in three-dimensional breast imaging for both screening and diagnostic purposes. It exhibits enhanced performance in identifying breast abnormalities and generates images of superior diagnostic quality with improved precision compared to conventional digital mammography. This improvement is particularly notable when examining dense breast tissue, which has traditionally been one of the most challenging aspects of breast imaging.

Increasing Wear Resistance of Ploughshare by Electrofriction Treatment

Surface layers of agricultural machinery working bodies are subjected to intensive abrasive wear during operation, which leads to rapid wear of equipment and reduction of its service life. To increase the wear resistance of the working surfaces of tools, the method of induction cladding using ‘Sormait-1’ materials is widely used. However, after coating, additional heat treatment is required, which improves physical and mechanical properties of the material and increases its durability. When using electrofriction technology (EFT) hardening, the surface of the parts is subjected to melting under the influence of electric arcs, which affects the surface characteristics of the coatings. In this work, two types of surface treatment of L53 steel were investigated: induction cladding using ‘Sormait-1’ material, as well as a combination of induction cladding and subsequent electrofriction treatment. The coatings were characterized and compared with the substrate in terms of the following parameters: microstructure, phase composition, hardness distribution, and friction-wear characteristics. After induction cladding of the Sormait-1 material, a dendritic structure was formed; however, subsequent electrofriction treatment resulted in a reduction of this dendritic structure, which contributed to an increase in the hardness of the material. The average hardness of the coatings after electrofriction treatment was 786 HV0.1, which is more than three times the hardness of the substrate. Furthermore, the influence of structural characteristics and hardness on abrasive wear resistance was examined in accordance with ASTM G65 international standards. Field tests were conducted on plough shares before and after electrofriction hardening to evaluate their performance. Each ploughshare was scanned with a structured 3D scanner before and after use in the field. From the scan data, the cutting-edge profile was calculated and three key parameters were determined: linear wear, volumetric wear, and mass reduction. According to the results of field tests, it was found that the service life of the blades hardened by electrofriction technology was 12%–14% higher compared to serial blades processed by induction cladding with the use of ‘Sormait-1’ material. Operational tests of hardened plough shares confirmed the results of laboratory tests and proved the advantages of electrofriction technology for increasing the wear resistance of soil tillage machine working bodies.

Analysis of the indoor wind environment in buildings on the Qinghai-Tibet plateau of China: A case study of the Dege Scripture Printing House

The Dege Scripture Printing House (DSPH), built in 1729, is located on the southeastern edge of the Qinghai-Tibet plateau and serves as a significant center for Tibetan Buddhism culture. Compared to typical Tibetan architecture, DSPH features unique characteristics, such as its substantial size, diverse structures, and adaptability to the local climate. The surrounding topography of DSPH primarily comprises plateaus and hilly plains, and the region experiences a continental plateau monsoon climate. This climate is characterized by cold temperatures due to high altitude, low precipitation, long sunshine hours, dry air, long winters, short summers, and intense solar radiation. To thoroughly assess the ventilation status of DSPH, a 3D scan of the building, physical environment tests, and indoor wind environment simulations were carried out. The field test results show that the unique architectural design of DSPH helps mitigate the impact of the harsh climate on indoor activities. By analyzing and evaluating the wind environment of DSPH, locally adapted climate resilience practices are summarized, and the physical properties of buildings characteristic of the Tibetan plateau are clarified. These findings provide a reference for promoting the development of local climate-adapted buildings, preserving traditional architectural culture, and promoting the integration of modern architecture with traditional culture. Furthermore, this study represents an important step toward enhancing energy efficiency and green building practices in ethnic minority settlements, contributing to sustainable development in the region.

3D Assessment of Mandibular Margin Morphological Change after BTX-A Injection for Masseter Hypertrophy: A Retrospective Study.

Some facial morphological changes may occur in some patients after botulinum toxin A (BTX-A) injection for the treatment of masseteric hypertrophy. The aim of this study was to investigate the morphological changes that occurred on mandibular margin. This retrospective analysis included 43 patients (86 lateral cases) who received BTX-A injections for masseter hypertrophy. To assess the morphological changes in the mandibular margin after treatment, the sharpness of the mandibular margin was evaluated before and after the injections using an established grading system. Additionally, 3D scanning technology was used to measure both the therapeutic effect of the BTX-A injections and the extent of soft tissue sagging in the mandibular margin area. Grading: The grade of mandibular margin sharpness change occurred in 41 lateral cases post-procedurally. Based on the measurement results, statistically significant differences were observed before, 1 month, and 3 months post-procedurally in subjects' positive lower face width representing the therapeutic effect of BTX-A injection and the extent of sagging soft tissue at sagittal section B and section C (p < 0.05), but not in section A (A: outer canthus, B: midpoint of the eyelid margin, and C: inner canthus) (p > 0.05). After receiving a BTX-A injection for masseteric hypertrophy, the mandibular margin may develop an unclear mandibular margin and sagging soft tissue in mandibular margin area. EBM Level: IV LEVEL OF EVIDENCE IV: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266.

On efficient 3D object retrieval

Due to the growth of the 3D technology, digital 3D models represented in the form of point clouds have attracted a lot of attention from both industry and academia. In this paper, due to a variety of applications, we study a fundamental problem called the 3D object retrieval, which is to find a set of 3D point clouds stored in a database that are similar to a given query 3D point cloud. To the best of our knowledge, solving the problem of 3D object retrieval efficiently remains unexplored in the research community. In this paper, we propose a framework called C2O to find the answer efficiently with the help of an index built on the database. In most of our experiments, C2O performs up to 2 orders of magnitude faster than all adapted algorithms in the literature. In particular, when the database size scales up to 100 million points, C2O answers the 3D object retrieval within 10 s but all adapted exact algorithms need more than 1000 s.

Reduced dose helical CT scout imaging on next generation wide volume CT system decreases scan length and overall radiation exposure

PurposeTraditional CT acquisition planning is based on scout projection images from planar anterior-posterior and lateral projections where the radiographer estimates organ locations. Alternatively, a new scout method utilizing ultra-low dose helical CT (3D Landmark Scan) offers cross-sectional imaging to identify anatomic structures in conjunction with artificial intelligence based Anatomic Landmark Detection (ALD) for automatic CT acquisition planning. The purpose of this study is to quantify changes in scan length and radiation dose of CT examinations planned using 3D Landmark Scan and ALD and performed on next generation wide volume CT versus examinations planned using traditional scout methods. We additionally aim to quantify changes in radiation dose reduction of scans planned with 3D Landmark Scan and performed on next generation wide volume CT. MethodsSingle-center retrospective analysis of consecutive patients with prior CT scan of the same organ who underwent clinical CT using 3D Landmark Scan and automatic scan planning. Acquisition length and dose-length-product (DLP) were collected. Data was analyzed by paired t-tests. Results104 total CT examinations (48.1 % chest, 15.4 % abdomen, 36.5 % chest/abdomen/pelvis) on 61 individual consecutive patients at a single center were retrospectively analyzed. 79.8 % of scans using 3D Landmark Scan had reduction in acquisition length compared to the respective prior acquisition. Median acquisition length using 3D Landmark Scan was 26.7 mm shorter than that using traditional scout methods (p < 0.001) with a 23.3 % median total radiation dose reduction (245.6 (IQR 150.0–400.8) mGy cm vs 320.3 (IQR 184.1–547.9) mGy cm). CT dose index similarly was overall decreased for scans planned with 3D Landmark and ALD and performed on next generation CT versus traditional methods (4.85 (IQR 3.8–7) mGy vs. 6.70 (IQR 4.43–9.18) mGy, respectively, p < 0.001). ConclusionScout imaging using reduced dose 3D Landmark Scan images and Anatomic Landmark Detection reduces acquisition range in chest, abdomen, and chest/abdomen/pelvis CT scans. This technology, in combination with next generation wide volume CT reduces total radiation dose.

Surface feature extraction method for cloud data of aircraft wall panel measurement points

In the cloud, users need to connect to the data server to perform the file transmission via the Internet, and the Server transmits data to many servers. A machine or vehicle that can fly with the assistance of the air is known as an Aircraft. As an alternative to the downward thrust of jet engines, it uses either static lift or an airfoil's dynamic lift to combat gravity's pull. Drawing wall panel measurement points in the model is easy using the Aircraft Wall Panels (AWP) button. Draw wall panels between existing nodes or on the drawing grid using the relevant wall panel specifications. The technique intends to discover and extract information about undesirable defects such as dents, protrusions, or scratches based on local surface attributes gathered from a 3D scanner. Defects from a perfectly smooth surface include indentations and bumps on the surface. An image's features may be extracted by reducing the number of pixels in the picture to a manageable size so that the most exciting sections of the image can be recorded with Surface Feature Extraction (SFE). Some of the problems are the threat of drones and composite materials that do not break easily in oxymoronic. The aircraft's inner structure may have been damaged, although this is impossible to determine. A runway incursion severely threatens aviation safety because of the rise in aircraft movement on the airport surface and other human factors. An electronic moving map of airport runways and taxiways is shown to the pilot through a head-up display in the cockpit's head-down position. A practical feature extraction approach is required to ensure the safety of the airport scene in runway incursion prevention systems. All the drawbacks are rectified by AWP-SFE sensors installed along the runway centerline to detect magnetic signals generated by surface-moving targets, and this information is utilized to compute the target's length. The target length may extract peak features after regularizing the time domain data. Differentiation of target characteristics is used to determine the similarities between distinct targets. The suggested method's signal characteristics are more easily recognized than time domain or frequency domain feature methods. The experimental results show the proposed method AWP-SE to achieve a high-efficiency ratio of 88.2%, activity ratio of 73.3%, Analysis of aircraft in wall plane measurement point of 87.8% and an error rate of 32.3% compared to other methods.

Understanding Occlusion and Temporomandibular Joint Function Using Deep Learning and Predictive Modeling.

Advancements in artificial intelligence (AI)-driven predictive modeling in dentistry are outpacing the clinical translation of research findings. Predictive modeling uses statistical methods to anticipate norms related to TMJ dynamics, complementing imaging modalities like cone beam computed tomography (CBCT) and magnetic resonance imaging (MRI). Deep learning, a subset of AI, helps quantify and analyze complex hierarchical relationships in occlusion and TMJ function. This narrative review explores the application of predictive modeling and deep learning to identify clinical trends and associations related to occlusion and TMJ function. Debates persist regarding best practices for managing occlusal factors in temporomandibular joint (TMJ) function analysis while interpreting and quantifying findings related to the TMJ and occlusion and mitigating biases remain challenging. Data generated from noninvasive chairside tools such as jaw trackers, video tracking, and 3D scanners with virtual articulators offer unique insights by predicting variations in dynamic jaw movement, TMJ, and occlusion. The predictions help us understand the highly individualized norms surrounding TMJ function that are often required to address temporomandibular disorders (TMDs) in general practice. Normal TMJ function, occlusion, and the appropriate management of TMDs are complex and continue to attract ongoing debate. This review examines how predictive modeling and artificial intelligence aid in understanding occlusion and TMJ function and provides insights into complex dental conditions such as TMDs that may improve diagnosis and treatment outcomes with noninvasive techniques.

GAFAR revisited—Exploring the limits of point cloud registration on sparse subsets

Robust estimation of a rigid transformation aligning two point clouds is a fundamental building block in many tasks in computer vision and robotics, such as mapping, (re-)localization, SLAM or 3D scanning, and many more. Deep-Learning based registration methods have proven superior in terms of robustness to outliers and bad initializations, yet their resource needs often render them unsuitable for applications where space, energy and real-time constraints come into play. Since - provided a good initialization - conventional registration algorithms like ICP prove to be fast and accurate, we may relax the requirements of absolute precision of a learning-based matcher and instead focus our attention on efficiency, robustness and generalization.In prior work, we introduced our small, fast and light-weight registration algorithm GAFAR, which works on sparse subsets of point clouds and exhibits a small enough footprint to be deployed on edge devices in mobile applications, while still providing accurate results and promising generalization ability. Based thereon, we develop this idea further towards applicability on real world data with improvements to the algorithm as well as further evaluations and analyses. We showcase this by applying it to Kitti Odometry Benchmark and 3DMatch dataset as well as demonstrating its usability on edge devices. The code and trained weights are published in https://github.com/mordecaimalignatius/GAFAR/.