The system, called MuHP-DLNet, was developed by Zhengchao Chang, Xianlin Peng, Jinye Peng, Shaohui Ma, Dong Liu, Dan Liu, and Moncef Gabbouj, a team spanning Northwest University and the Henan Institute of Technology in China, Henan Normal University, and Tampere University in Finland. Their work addresses a problem that has long frustrated automated heritage inspection: damage on murals is sparse, irregular, and visually entangled with the artwork itself. A crack in a painted drapery fold or a flake of lost pigment inside a textured cloud pattern is not a clean, isolated object sitting on a plain background. It is a small anomaly buried inside a richly decorated surface, which makes it extraordinarily hard for a segmentation network to separate damage from art.

This difficulty has a technical name: foreground imbalance. In a typical mural photograph, damaged pixels make up only a tiny fraction of the image, while intact painted surface dominates. Standard segmentation networks, which learn by minimizing errors across all pixels equally, are quickly overwhelmed by the intact majority. They drift toward predicting that nothing is damaged, because that safe guess produces low average error even though it misses exactly the information conservators need. The problem compounds at the boundaries of damaged regions, where irregular, jagged edges of flaking plaster resist the smooth contours that convolutional networks naturally prefer to draw.

MuHP-DLNet’s answer is a multi-stage architecture that splits the job in two. A coarse branch first performs global localization, scanning the whole image to identify roughly where damage is likely to occur. This stage acts like a conservator’s initial walkthrough of a hall: it does not trace every hairline crack, but it establishes which wall sections deserve close attention. A fine branch then takes over, performing detailed segmentation on the regions the coarse stage has flagged. This division of labor means the network does not have to simultaneously find damage and delineate it at pixel-level precision in a single pass, a combination that has historically forced compromises in accuracy.

Within the fine branch, three components work in concert. Edge enhancement sharpens the network’s sensitivity to the transitions between damaged and intact material, directly attacking the boundary segmentation problem. Transformer encoding then models long-range relationships across the image, allowing the network to reason about context that lies far from any given pixel. This matters because damage patterns are not random: deterioration often follows structural lines in the wall, and recognizing those global relationships helps the network distinguish a genuine crack from a painted line that merely looks like one. Finally, query-guided dynamic feature modulation lets the network adapt its internal feature processing to the specific characteristics of each damaged region, rather than applying a single rigid filter everywhere.

To fuse these two streams of information, the architecture includes cross-scale aggregation, which combines the global context captured by the coarse branch with the local details extracted by the fine branch. The goal is to improve mask coherence, meaning that the segmented damage regions come out as plausible, connected shapes rather than scattered fragments, and to sharpen boundary delineation so that the outline of each damaged area is traced accurately. For conservators, coherent masks matter enormously: a fragmented prediction forces manual cleanup, while a well-formed region map can feed directly into condition reports and restoration planning.

Perhaps the study’s most distinctive contribution lies not in the architecture but in the loss function, the mathematical signal that tells the network how wrong its predictions are during training. The team designed a damage-aware loss that combines weighted binary cross entropy with the Dice loss, a pairing common in medical image segmentation where small structures must be found in large images. The innovation is in how the weighting is computed: it uses the image’s damage ratio, the actual proportion of damaged pixels, together with the local structure of the annotations, so that the training signal adapts to each image’s specific condition. A mural that is nearly intact generates different weighting than one with extensive deterioration. Learnable normalized weights further balance the main supervision signal against auxiliary supervision from intermediate network stages, letting the training process discover the optimal emphasis on its own rather than relying on hand-tuned constants.

The experimental evaluation covered two mural datasets and three public benchmarks, where MuHP-DLNet demonstrated favorable overall performance against representative baseline methods. Just as importantly, the team ran controlled ablation studies, systematically removing individual components to measure each one’s contribution. These ablations identified complementary architectural contributions, confirming that the edge enhancement, Transformer encoding, dynamic modulation, and cross-scale aggregation each add distinct value. The ablation results also delivered a striking finding: the proposed damage-aware loss had the largest effect among all the evaluated components. In other words, teaching the network to weigh its mistakes intelligently mattered more than any single architectural refinement, a conclusion that resonates with a broader trend in computer vision toward smarter supervision rather than ever-larger models.

The practical implications extend across the heritage conservation workflow. Accurate damage maps are the foundation of condition assessment, prioritization of interventions, and monitoring of deterioration over time. When such maps can be produced automatically from ordinary photographs, conservators can survey large sites more frequently and consistently, tracking whether a crack is spreading or a flaking area is stabilizing. The method’s demonstrated ability to segment subtle damage within complex mural textures suggests it could handle the demanding visual conditions of real heritage sites, where painted ornamentation and deterioration often share colors, textures, and contours.

The work also fits into a rapidly growing research effort at the intersection of artificial intelligence and heritage science. Related recent publications in the same journal include a damage detection network based on multi-scale boundary and region feature fusion, an ancient mural classification model using shallow texture enhancement and cross-layer attention aggregation, and a mural inpainting approach built on a two-stage generative adversarial network. Together, these studies sketch a pipeline in which machine learning systems detect, classify, and even digitally repair deterioration. MuHP-DLNet’s contribution to that pipeline is a reminder that the hardest problems in automated heritage documentation are often not about seeing more, but about learning what to look for when the evidence is faint, irregular, and woven into the fabric of the artwork itself. The research was supported by the National Natural Science Foundation of China and the Key Technologies Research and Development Program of Henan Province, and the authors report no competing interests.

Subject of Research: Deep learning-based damage detection and segmentation in ancient mural conservation

Article Title: MuHP-DLNet: a multi-stage high-precision mural damage detection network with damage-aware dynamic loss

Article References: Chang, Z., Peng, X., Peng, J., Ma, S., Liu, D., Liu, D., & Gabbouj, M. (2026). MuHP-DLNet: a multi-stage high-precision mural damage detection network with damage-aware dynamic loss. npj Heritage Science. https://doi.org/10.1038/s40494-026-03016-2

Image Credits: AI Generated

DOI: 10.1038/s40494-026-03016-2

Keywords: ancient murals, deep learning, image segmentation, heritage science, damage detection, Transformer, loss function, computer vision, conservation, foreground imbalance, boundary delineation, npj Heritage Science

The research, led by Fan Sun, Qing Chun, Shiyu Ma and Yu Yuan of the School of Architecture at Southeast University in Nanjing, addresses a persistent bottleneck in heritage documentation. Image-based inspection methods, which have become popular because cameras are cheap and photographs are easy to collect, suffer from two fundamental weaknesses. First, they depend on illumination: shadows, glare and uneven lighting can hide damage or create false impressions of it. Second, a two-dimensional photograph loses depth information, so a shallow surface flake and a deep cavity that look identical in plan may be indistinguishable without laborious supplementary measurement. Point cloud data, captured by terrestrial laser scanning or photogrammetric reconstruction, sidestep both problems by recording dense, accurate three-dimensional geometry regardless of surface color or lighting conditions.

But raw point clouds bring their own challenge. Unlike an image, which arrives as a tidy grid of pixels, a point cloud is an unordered set of millions of XYZ coordinates with irregular spacing. Convolutional neural networks designed for photographs cannot be applied directly. The Southeast University team built their pipeline on PointNet++, a neural architecture specifically designed for point sets, which learns hierarchical features by grouping points into local neighborhoods and abstracting them layer by layer. Their contribution is an improved version of this architecture incorporating what they call a Novel Set Abstraction, or NSA, module, which refines how the network samples and aggregates geometric information from the scanned surface.

The framework operates in three integrated stages. The first is preprocessing, in which the raw scans are cleaned and prepared so that the network receives consistent input. The second is binary segmentation: the improved PointNet++ classifies every single point in the cloud as either damaged or undamaged, effectively painting a per-point diagnosis across the entire wall surface. The third stage is quantification, where morphological analysis of the segmented damage regions converts the classification into numbers that conservators actually need, namely the damaged area and the maximum depth of each deterioration zone. This last step exploits the very property that images lack, because depth is measured directly from the three-dimensional geometry of the scan.

To train and evaluate the model, the researchers assembled a dataset of 317 annotated patches collected from four wall segments at different locations and orientations along the Nanjing City Wall. This variety matters. A wall facing south weathers differently from one facing north, and damage near the base, where moisture and salt accumulate, differs from damage higher up. By sampling across locations and orientations, the dataset captures a representative range of the surface deterioration that the wall actually exhibits, giving the network a fair test of generalization within the structure it was designed to monitor.

The performance figures tell a striking story about the value of the architectural improvement. The original PointNet++ achieved an intersection over union, the standard overlap metric for segmentation tasks, of only 28.80 percent on the damage class, meaning the baseline network struggled to delineate eroded regions accurately. The improved model with the NSA module raised that figure to 55.98 percent, nearly doubling the damage-class score, and reached a mean intersection over union of 73.77 percent across all classes. In segmentation problems where damage boundaries are fuzzy and the damaged class occupies a minority of points, such gains are substantial rather than incremental.

Segmentation alone, however, does not answer the conservator’s question of how much wall has been lost. The team therefore validated the full pipeline, including the quantification stage, against 30 independently measured damaged regions. The automated framework produced a mean relative area error of 17.24 percent and a mean absolute maximum-depth error of just 6.77 millimeters. For a structure whose deterioration must be tracked over years and decades, an error of under seven millimeters in depth estimation, obtained without a single manual measurement, represents a meaningful advance in the precision of routine condition surveys.

The implications extend beyond one famous wall. Masonry architectural heritage is vast and heterogeneous, spanning city walls, temples, churches, fortifications and vernacular buildings on every continent, and the professionals responsible for it are chronically outnumbered by the fabric they must protect. A framework that turns laser scans, which many heritage agencies already collect for archival purposes, into quantitative damage maps could transform those archives from static records into active monitoring tools. Repeated scans of the same wall, processed by the same model, would allow deterioration rates to be compared over time, revealing which sections are stable and which are actively degrading, and helping prioritize expensive conservation interventions where they are needed most.

The authors are appropriately measured about the limits of their results. The framework was evaluated under the tested conditions of the Nanjing City Wall, and they note that it supports preliminary documentation and monitoring of surface material deterioration there, while its transferability to other structures and environments remains to be validated. That caution is scientifically sound. Deep learning models trained on one building’s materials, weathering patterns and scanning conditions cannot be assumed to perform equally on a sandstone cathedral in a different climate, and the researchers’ explicit acknowledgment of this boundary sets a clear agenda for future work on cross-site generalization.

Even with that caveat, the study demonstrates a convincing convergence of two technologies that heritage science has been pursuing in parallel for years. Three-dimensional scanning provides the geometric fidelity that photographs cannot, and deep learning provides the automation that manual annotation cannot scale to. By combining an improved point-set neural network with morphology-based quantification, the Southeast University team has shown that a wall can, in effect, report its own injuries: where the surface has failed, how extensive the failure is, and how deep it reaches. As such systems mature and are tested across more sites, the slow, invisible erosion of the world’s masonry heritage may become visible, measurable and, ultimately, far easier to manage.

Subject of Research: Deep learning-based detection and quantification of surface damage in masonry architectural heritage using 3D point cloud data

Article Title: A method for detecting and quantifying damage in masonry architectural heritage based on deep learning of point cloud data

Article References: Sun, F., Chun, Q., Ma, S., & Yuan, Y. (2026). A method for detecting and quantifying damage in masonry architectural heritage based on deep learning of point cloud data. npj Heritage Science. https://doi.org/10.1038/s40494-026-03043-z

Image Credits: AI Generated

DOI: 10.1038/s40494-026-03043-z

Keywords: deep learning, point cloud, masonry heritage, Nanjing City Wall, PointNet++, damage detection, 3D segmentation, heritage conservation, laser scanning, surface deterioration, structural health monitoring, npj Heritage Science

The new framework, described in the journal Smart Agricultural Technology, tackles a problem that has long frustrated agricultural engineers. Sugar beet is one of Europe’s most valuable arable crops, with France and Germany leading production according to the United Nations Food and Agriculture Organization, yet a meaningful share of every harvest is degraded by mechanical damage between the soil and the truck. Because the damage happens inside a machine moving at full field speed, farmers have had no practical way to see it, measure it, or respond to it. The researchers, led by Konstantinos Gkountakos of the Centre for Research and Technology Hellas, set out to give harvesters something close to a nervous system: cameras, neural networks, and the ability to distinguish a healthy beet from a broken one in real time.

The journey from field to tank inside a modern harvester is short but violent. First, defoliators gently remove the leafy tops; cut too high and leaves remain to regrow, cut too low and the head of the beet, which is rich in sugar, is shaved away. Next, lifting shares pry the roots from the ground, ideally without cracking them. Then come the cleaning turbines, cylindrical drums that spin the beets against one another to knock off soil, stones, and debris. Residual soil inflates transport costs and accelerates erosion, but overly aggressive cleaning shatters and abrades the roots. Finally, a metallic sieve belt carries the cleaned beets into the tank. At the end of this conveyor, just before loading, the researchers saw an opportunity: a natural inspection point where every beet passes the camera, and where a damaged crop could still trigger adjustments to the lifting and cleaning mechanisms upstream.

To train and test their system, the team created something the field has lacked: a genuinely realistic dataset, which they named VerBeet. Using an industrial Basler camera with 8.3 megapixel resolution mounted in a white housing above the sieve belt of a Vervaet Q-621 harvester, they captured 694 frames over five harvesting days between early December 2025 and early January 2026, spaced 60 to 75 seconds apart to avoid overlapping content and to sample entire fields. Because daylight shifted with the harvester’s position relative to the sun, producing frames that ranged from dark to glare-bright, they installed three 40-watt LED lights on magnetic mounts, providing 13,200 lumens of consistent illumination. Two human annotators then spent roughly 17 hours drawing bounding boxes around every beet, cross-validating each other’s work, ultimately labeling 5,158 individual beets across three categories: undamaged, breakage, and abrasion.

The class distribution itself tells a story about mechanized harvesting. As expected, undamaged beets dominate with 3,685 instances, but 1,293 beets showed breakage, structural damage to the root from lifting or cleaning, and 180 displayed surface abrasion. The imbalance between the two damage types proved instructive: abrasions, the researchers found, are often hidden by clinging soil or simply invisible from a single camera angle, whereas a broken beet announces itself regardless of perspective. This matters because the ultimate goal is not just detection but attribution, knowing whether damage originated in the topping, lifting, or cleaning stages, so the harvester can adjust the right component rather than guessing.

The framework itself operates in two stages. A state-of-the-art object detector first localizes every beet in each frame; the researchers benchmarked lightweight members of the YOLO family, including YOLOv11n, YOLOv12n, and the newest YOLO26, against the transformer-based RT-DETR. Each cropped beet is then passed to a convolutional neural network classifier, drawn from EfficientNet, DenseNet, and MobileNet architectures, that decides whether the root is whole, broken, or abraded. Speed was non-negotiable: YOLOv11n processed more than 160 frames per second on the VerBeet test set, while MobileNetV2 classified beets at around 130 instances per second on modest consumer-grade hardware, confirming the pipeline could run on the edge devices that a real harvester would carry.

Two clever training tricks pushed performance further. The first is a size-aware weighted loss. Because the camera always views the belt from the same distance, a beet’s apparent size in the image reveals how much of it is visible; occluded or partially buried beets appear smaller and are harder to classify. The loss function weights each training sample by the inverse of its normalized area, so smaller, harder instances contribute more to learning, with weights clipped between 0.1 and 1.0 to keep training stable. The second trick is generative: the team used FLUX.2-4B Klein, a lightweight open-source diffusion model, to synthesize augmented training images in which each beet is placed on a randomized background. This counteracts overfitting to the unchanging metallic belt behind the beets and, crucially, multiplies scarce examples of the rare abrasion class. Every synthetic image was manually inspected, and hallucinated samples where the model invented or erased beets were discarded.

The results reveal both the promise and the stubborn difficulty of real-world machine vision. On the controlled, laboratory-style Semantic Sugar Beet dataset, models reached F1-scores above 87 percent, but on VerBeet’s dusty, occluded, unevenly lit frames, the best binary damage detection, EfficientNetB3 with size-aware loss and diffusion augmentation, achieved an F1-score of 71.48 percent. Cross-dataset experiments drove the point home: models trained on the controlled dataset collapsed when tested in the field, with recall as low as 12.7 percent, while models trained on VerBeet generalized far better, reaching 74.6 percent mAP on the cleaner data. In the most demanding test, a field-disjoint end-to-end evaluation where the test set always contained beets from a harvesting day entirely unseen during training, the full pipeline of YOLOv11n and MobileNetV2-Size achieved a mean F1-score of 67.09 percent across five folds, distinguishing beets, breakage, and abrasion in realistic conditions.

The researchers are candid about limitations. The dataset came from a single harvester model, so cameras mounted at different angles or heights on other machines could shift the statistics the networks rely on. Abrasion remains the weakest link, with multiclass F1-scores hovering near 51 percent, limited by the scarcity of genuine examples and the visual similarity between abraded and broken surfaces. Future work, the authors suggest, will extend VerBeet to more harvesting conditions, beet varieties, and even other root crops such as potatoes, and will explore adapting modern vision foundation models to withstand occlusion, dust, and erratic light under real-time constraints on semi-autonomous harvesting platforms.

Still, the trajectory is clear. Agriculture is racing toward machines that perceive, decide, and react, and the sugar beet harvester, long a blunt instrument of force and throughput, is acquiring finesse. A future machine that detects rising breakage rates and automatically softens its lifting shares, or notices abraded beets and eases the turbine speed, would translate directly into tonnes of sugar saved and soil left in the field where it belongs. The VerBeet dataset, with its 5,158 honestly annotated beets captured in the mud and glare of actual harvest days, may well become the benchmark on which that generation of gentle machines is trained.

Subject of Research: Real-time computer vision detection of sugar beet damage during mechanical harvesting

Article Title: Sugar beet localization and damage detection during harvesting

Article References: Gkountakos, K., Pasios, S., Ioannidis, K., Demestichas, K., Vrochidis, S., & Kompatsiaris, I. (2026). Sugar beet localization and damage detection during harvesting. Smart Agricultural Technology, 15, Article 102560. https://doi.org/10.1016/j.atech.2026.102560

Image Credits: AI Generated

DOI: 10.1016/j.atech.2026.102560

Keywords: sugar beet, computer vision, deep learning, YOLO, damage detection, precision agriculture, harvesting, VerBeet dataset, diffusion models, object detection, agricultural technology, neural networks