Alper Yilmaz

Shehan Perera, Yunus Erzurumlu, Deepak Gulati et al.

Skin cancer segmentation poses a significant challenge in medical image analysis. Numerous existing solutions, predominantly CNN-based, face issues related to a lack of global contextual understanding. Alternatively, some approaches resort to large-scale Transformer models to bridge the global contextual gaps, but at the expense of model size and computational complexity. Finally many Transformer based approaches rely primarily on CNN based decoders overlooking the benefits of Transformer based decoding models. Recognizing these limitations, we address the need efficient lightweight solutions by introducing MobileUNETR, which aims to overcome the performance constraints associated with both CNNs and Transformers while minimizing model size, presenting a promising stride towards efficient image segmentation. MobileUNETR has 3 main features. 1) MobileUNETR comprises of a lightweight hybrid CNN-Transformer encoder to help balance local and global contextual feature extraction in an efficient manner; 2) A novel hybrid decoder that simultaneously utilizes low-level and global features at different resolutions within the decoding stage for accurate mask generation; 3) surpassing large and complex architectures, MobileUNETR achieves superior performance with 3 million parameters and a computational complexity of 1.3 GFLOP resulting in 10x and 23x reduction in parameters and FLOPS, respectively. Extensive experiments have been conducted to validate the effectiveness of our proposed method on four publicly available skin lesion segmentation datasets, including ISIC 2016, ISIC 2017, ISIC 2018, and PH2 datasets. The code will be publicly available at: https://github.com/OSUPCVLab/MobileUNETR.git

Yongsheng Bai, Bing Zha, Halil Sezen et al.

This paper presents a few comprehensive experimental studies for automated Structural Damage Detection (SDD) in extreme events using deep learning methods for processing 2D images. In the first study, a 152-layer Residual network (ResNet) is utilized to classify multiple classes in eight SDD tasks, which include identification of scene levels, damage levels, material types, etc. The proposed ResNet achieved high accuracy for each task while the positions of the damage are not identifiable. In the second study, the existing ResNet and a segmentation network (U-Net) are combined into a new pipeline, cascaded networks, for categorizing and locating structural damage. The results show that the accuracy of damage detection is significantly improved compared to only using a segmentation network. In the third and fourth studies, end-to-end networks are developed and tested as a new solution to directly detect cracks and spalling in the image collections of recent large earthquakes. One of the proposed networks can achieve an accuracy above 67.6% for all tested images at various scales and resolutions, and shows its robustness for these human-free detection tasks. As a preliminary field study, we applied the proposed method to detect damage in a concrete structure that was tested to study its progressive collapse performance. The experiments indicate that these solutions for automatic detection of structural damage using deep learning methods are feasible and promising. The training datasets and codes will be made available for the public upon the publication of this paper.

Jianli Wei, Deniz Karakay, Alper Yilmaz

The Global Positioning System (GPS) has become a part of our daily life with the primary goal of providing geopositioning service. For an unmanned aerial system (UAS), geolocalization ability is an extremely important necessity which is achieved using Inertial Navigation System (INS) with the GPS at its heart. Without geopositioning service, UAS is unable to fly to its destination or come back home. Unfortunately, GPS signals can be jammed and suffer from a multipath problem in urban canyons. Our goal is to propose an alternative approach to geolocalize a UAS when GPS signal is degraded or denied. Considering UAS has a downward-looking camera on its platform that can acquire real-time images as the platform flies, we apply modern deep learning techniques to achieve geolocalization. In particular, we perform image matching to establish latent feature conjugates between UAS acquired imagery and satellite orthophotos. A typical application of feature matching suffers from high-rise buildings and new constructions in the field that introduce uncertainties into homography estimation, hence results in poor geolocalization performance. Instead, we extract GIS information from OpenStreetMap (OSM) to semantically segment matched features into building and terrain classes. The GIS mask works as a filter in selecting semantically matched features that enhance coplanarity conditions and the UAS geolocalization accuracy. Once the paper is published our code will be publicly available at https://github.com/OSUPCVLab/UbihereDrone2021.

Alexys Herleym Rodríguez Avellaneda, Abdollah Shafieezadeh, Alper Yilmaz

This paper investigates whether socioeconomic factors are important for the hurricane performance of the electric power system in Florida. The investigation is performed using the Random Forest classifier with Mean Decrease of Accuracy (MDA) for measuring the importance of a set of factors that include hazard intensity, time to recovery from maximum impact, and socioeconomic characteristics of the affected population. The data set (at county scale) for this study includes socioeconomic variables from the 5-year American Community Survey (ACS), as well as wind velocities, and outage data of five hurricanes including Alberto and Michael in 2018, Dorian in 2019, and Eta and Isaias in 2020. The study shows that socioeconomic variables are considerably important for the system performance model. This indicates that social disparities may exist in the occurrence of power outages, which directly impact the resilience of communities and thus require immediate attention.

Yuci Han, Jianli Wei, Alper Yilmaz

This paper presents a novel end-to-end Unmanned Aerial System (UAS) navigation approach for long-range visual navigation in the real world. Inspired by dual-process visual navigation system of human's instinct: environment understanding and landmark recognition, we formulate the UAS navigation task into two same phases. Our system combines the reinforcement learning (RL) and image matching approaches. First, the agent learns the navigation policy using RL in the specified environment. To achieve this, we design an interactive UASNAV environment for the training process. Once the agent learns the navigation policy, which means 'familiarized themselves with the environment', we let the UAS fly in the real world to recognize the landmarks using image matching method and take action according to the learned policy. During the navigation process, the UAS is embedded with single camera as the only visual sensor. We demonstrate that the UAS can learn navigating to the destination hundreds meters away from the starting point with the shortest path in the real world scenario.

Yuci Han, Alper Yilmaz

In this paper, we propose Sparse Imitation Reinforcement Learning (SIRL), a hybrid end-to-end control policy that combines the sparse expert driving knowledge with reinforcement learning (RL) policy for autonomous driving (AD) task in CARLA simulation environment. The sparse expert is designed based on hand-crafted rules which is suboptimal but provides a risk-averse strategy by enforcing experience for critical scenarios such as pedestrian and vehicle avoidance, and traffic light detection. As it has been demonstrated, training a RL agent from scratch is data-inefficient and time consuming particularly for the urban driving task, due to the complexity of situations stemming from the vast size of state space. Our SIRL strategy provides a solution to solve these problems by fusing the output distribution of the sparse expert policy and the RL policy to generate a composite driving policy. With the guidance of the sparse expert during the early training stage, SIRL strategy accelerates the training process and keeps the RL exploration from causing a catastrophe outcome, and ensures safe exploration. To some extent, the SIRL agent is imitating the driving expert's behavior. At the same time, it continuously gains knowledge during training therefore it keeps making improvement beyond the sparse expert, and can surpass both the sparse expert and a traditional RL agent. We experimentally validate the efficacy of proposed SIRL approach in a complex urban scenario within the CARLA simulator. Besides, we compare the SIRL agent's performance for risk-averse exploration and high learning efficiency with the traditional RL approach. We additionally demonstrate the SIRL agent's generalization ability to transfer the driving skill to unseen environment.

Shehan Perera, Pouyan Navard, Alper Yilmaz

The adoption of Vision Transformers (ViTs) based architectures represents a significant advancement in 3D Medical Image (MI) segmentation, surpassing traditional Convolutional Neural Network (CNN) models by enhancing global contextual understanding. While this paradigm shift has significantly enhanced 3D segmentation performance, state-of-the-art architectures require extremely large and complex architectures with large scale computing resources for training and deployment. Furthermore, in the context of limited datasets, often encountered in medical imaging, larger models can present hurdles in both model generalization and convergence. In response to these challenges and to demonstrate that lightweight models are a valuable area of research in 3D medical imaging, we present SegFormer3D, a hierarchical Transformer that calculates attention across multiscale volumetric features. Additionally, SegFormer3D avoids complex decoders and uses an all-MLP decoder to aggregate local and global attention features to produce highly accurate segmentation masks. The proposed memory efficient Transformer preserves the performance characteristics of a significantly larger model in a compact design. SegFormer3D democratizes deep learning for 3D medical image segmentation by offering a model with 33x less parameters and a 13x reduction in GFLOPS compared to the current state-of-the-art (SOTA). We benchmark SegFormer3D against the current SOTA models on three widely used datasets Synapse, BRaTs, and ACDC, achieving competitive results. Code: https://github.com/OSUPCVLab/SegFormer3D.git

Gokce Inal, Pouyan Navard, Alper Yilmaz

Recent advances in multimodal vision-language models (VLMs) have enabled joint reasoning over visual and textual information, yet their application to planetary science remains largely unexplored. A key hindrance is the absence of large-scale datasets that pair real planetary imagery with detailed scientific descriptions. In this work, we introduce LLaVA-LE (Large Language-and-Vision Assistant for Lunar Exploration), a vision-language model specialized for lunar surface and subsurface characterization. To enable this capability, we curate a new large-scale multimodal lunar dataset, LUCID (LUnar Caption Image Dataset) consisting of 96k high-resolution panchromatic images paired with detailed captions describing lunar terrain characteristics, and 81k question-answer (QA) pairs derived from approximately 20k images in the LUCID dataset. Leveraging this dataset, we fine-tune LLaVA using a two-stage training curriculum: (1) concept alignment for domain-specific terrain description, and (2) instruction-tuned visual question answering. We further design evaluation benchmarks spanning multiple levels of reasoning complexity relevant to lunar terrain analysis. Evaluated against GPT and Gemini judges, LLaVA-LE achieves a 3.3x overall performance gain over Base LLaVA and 2.1x over our Stage 1 model, with a reasoning score of 1.070, exceeding the judge's own reference score, highlighting the effectiveness of domain-specific multimodal data and instruction tuning to advance VLMs in planetary exploration. Code is available at https://github.com/OSUPCVLab/LLaVA-LE.

Michael Karnes, Alper Yilmaz

The field of deep visual recognition is undergoing a paradigm shift toward universal representations. The Platonic Representation Hypothesis suggests that diverse architectures trained on massive datasets are converging toward a shared, "ideal" latent space. This again raises a critical question: is a "Good Embedding All You Need?" In this paper, we leverage this convergence to demonstrate that off-the-shelf embeddings are inherently "good enough" for complex tasks, rendering intensive task-specific fine-tuning unnecessary. We explore this hypothesis within the few-shot learning framework, proposing a straightforward, non-parametric pipeline that entirely bypasses backpropagation. By utilizing a k-Nearest Neighbor classifier on frozen DINOv2-L features, we conduct a layer-wise characterization to identify an optimal feature extraction. We further demonstrate that manifold refinement via PCA and ICA provides a beneficial regularizing effect. Our results across four major benchmarks demonstrate that our approach consistently surpasses sophisticated meta-learning algorithms, achieving state-of-the-art performance.

Rahul Medicharla, Alper Yilmaz

In this paper, we introduce MotivNet, a generalizable facial emotion recognition model for robust real-world application. Current state-of-the-art FER models tend to have weak generalization when tested on diverse data, leading to deteriorated performance in the real world and hindering FER as a research domain. Though researchers have proposed complex architectures to address this generalization issue, they require training cross-domain to obtain generalizable results, which is inherently contradictory for real-world application. Our model, MotivNet, achieves competitive performance across datasets without cross-domain training by using Meta-Sapiens as a backbone. Sapiens is a human vision foundational model with state-of-the-art generalization in the real world through large-scale pretraining of a Masked Autoencoder. We propose MotivNet as an additional downstream task for Sapiens and define three criteria to evaluate MotivNet's viability as a Sapiens task: benchmark performance, model similarity, and data similarity. Throughout this paper, we describe the components of MotivNet, our training approach, and our results showing MotivNet is generalizable across domains. We demonstrate that MotivNet can be benchmarked against existing SOTA models and meets the listed criteria, validating MotivNet as a Sapiens downstream task, and making FER more incentivizing for in-the-wild application. The code is available at https://github.com/OSUPCVLab/EmotionFromFaceImages.

Yusen Peng, Alper Yilmaz

Skeleton-based human action recognition leverages sequences of human joint coordinates to identify actions performed in videos. Owing to the intrinsic spatiotemporal structure of skeleton data, Graph Convolutional Networks (GCNs) have been the dominant architecture in this field. However, recent advances in transformer models and masked pretraining frameworks open new avenues for representation learning. In this work, we propose CascadeFormer, a family of two-stage cascading transformers for skeleton-based human action recognition. Our framework consists of a masked pretraining stage to learn generalizable skeleton representations, followed by a cascading fine-tuning stage tailored for discriminative action classification. We evaluate CascadeFormer across three benchmark datasets (Penn Action N-UCLA, and NTU RGB+D 60), achieving competitive performance on all tasks. To promote reproducibility, we release our code and model checkpoints.

Michael Karnes, Shehan Perera, Srikar Adhikari et al.

This paper presents a novel ultrasound imaging point-of-care (PoC) COVID-19 diagnostic system. The adaptive visual diagnostics utilize few-shot learning (FSL) to generate encoded disease state models that are stored and classified using a dictionary of knowns. The novel vocabulary based feature processing of the pipeline adapts the knowledge of a pretrained deep neural network to compress the ultrasound images into discrimative descriptions. The computational efficiency of the FSL approach enables high diagnostic deep learning performance in PoC settings, where training data is limited and the annotation process is not strictly controlled. The algorithm performance is evaluated on the open source COVID-19 POCUS Dataset to validate the system's ability to distinguish COVID-19, pneumonia, and healthy disease states. The results of the empirical analyses demonstrate the appropriate efficiency and accuracy for scalable PoC use. The code for this work will be made publicly available on GitHub upon acceptance.

Jianli Wei, Guanyu Xu, Alper Yilmaz

Geopositioning and tracking a moving boat at sea is a very challenging problem, requiring boat detection, matching and estimating its GPS location from imagery with no common features. The problem can be stated as follows: given imagery from a camera mounted on a moving platform with known GPS location as the only valid sensor, we predict the geoposition of a target boat visible in images. Our solution uses recent ML algorithms, the camera-scene geometry and Bayesian filtering. The proposed pipeline first detects and tracks the target boat's location in the image with the strategy of tracking by detection. This image location is then converted to geoposition to the local sea coordinates referenced to the camera GPS location using plane projective geometry. Finally, target boat local coordinates are transformed to global GPS coordinates to estimate the geoposition. To achieve a smooth geotrajectory, we apply unscented Kalman filter (UKF) which implicitly overcomes small detection errors in the early stages of the pipeline. We tested the performance of our approach using GPS ground truth and show the accuracy and speed of the estimated geopositions. Our code is publicly available at https://github.com/JianliWei1995/AI-Track-at-Sea.

Jianzhu Huai, Yujia Zhang, Alper Yilmaz

In recent years, commodity mobile devices equipped with cameras and inertial measurement units (IMUs) have attracted much research and design effort for augmented reality (AR) and robotics applications. Based on such sensors, many commercial AR toolkits and public benchmark datasets have been made available to accelerate hatching and validating new ideas. To lower the difficulty and enhance the flexibility in accessing the rich raw data of typical AR sensors on mobile devices, this paper present the mobile AR sensor (MARS) logger for two of the most popular mobile operating systems, Android and iOS. The logger highlights the best possible synchronization between the camera and the IMU allowed by a mobile device, and efficient saving of images at about 30Hz, and recording the metadata relevant to AR applications. This logger has been tested on a relatively large spectrum of mobile devices, and the collected data has been used for analyzing the sensor characteristics. We see that this application will facilitate research and development related to AR and robotics, so it has been open sourced at https://github.com/OSUPCVLab/mobile-ar-sensor-logger.

Luming Yang, Haoxian Liu, Siqing Li et al.

Fine-grained action evaluation in medical vision faces unique challenges due to the unavailability of comprehensive datasets, stringent precision requirements, and insufficient spatiotemporal dynamic modeling of very rapid actions. To support development and evaluation, we introduce CPREval-6k, a multi-view, multi-label medical action benchmark containing 6,372 expert-annotated videos with 22 clinical labels. Using this dataset, we present GaussMedAct, a multivariate Gaussian encoding framework, to advance medical motion analysis through adaptive spatiotemporal representation learning. Multivariate Gaussian Representation projects the joint motions to a temporally scaled multi-dimensional space, and decomposes actions into adaptive 3D Gaussians that serve as tokens. These tokens preserve motion semantics through anisotropic covariance modeling while maintaining robustness to spatiotemporal noise. Hybrid Spatial Encoding, employing a Cartesian and Vector dual-stream strategy, effectively utilizes skeletal information in the form of joint and bone features. The proposed method achieves 92.1% Top-1 accuracy with real-time inference on the benchmark, outperforming the ST-GCN baseline by +5.9% accuracy with only 10% FLOPs. Cross-dataset experiments confirm the superiority of our method in robustness.

Yuci Han, Alper Yilmaz

This paper presents the World-Action Model (WAM), an action-regularized world model that jointly reasons over future visual observations and the actions that drive state transitions. Unlike conventional world models trained solely via image prediction, WAM incorporates an inverse dynamics objective into DreamerV2 that predicts actions from latent state transitions, encouraging the learned representations to capture action-relevant structure critical for downstream control. We evaluate WAM on enhancing policy learning across eight manipulation tasks from the CALVIN benchmark. We first pretrain a diffusion policy via behavioral cloning on world model latents, then refine it with model-based PPO inside the frozen world model. Without modifying the policy architecture or training procedure, WAM improves average behavioral cloning success from 59.4% to 71.2% over DreamerV2 and DiWA baselines. After PPO fine-tuning, WAM achieves 92.8% average success versus 79.8% for the baseline, with two tasks reaching 100%, using 8.7x fewer training steps.

Yunus Talha Erzurumlu, John E. Anderson, William J. Shuart et al.

Cross-view geolocalization (CVGL) systems, while effective at retrieving a list of relevant candidates (high Recall@k), often fail to identify the single best match (low Top-1 accuracy). This work investigates the use of zero-shot Vision-Language Models (VLMs) as rerankers to address this gap. We propose a two-stage framework: state-of-the-art (SOTA) retrieval followed by VLM reranking. We systematically compare two strategies: (1) Pointwise (scoring candidates individually) and (2) Pairwise (comparing candidates relatively). Experiments on the VIGOR dataset show a clear divergence: all pointwise methods cause a catastrophic drop in performance or no change at all. In contrast, a pairwise comparison strategy using LLaVA improves Top-1 accuracy over the strong retrieval baseline. Our analysis concludes that, these VLMs are poorly calibrated for absolute relevance scoring but are effective at fine-grained relative visual judgment, making pairwise reranking a promising direction for enhancing CVGL precision.

Yuci Han, Charles Toth, John E. Anderson et al.

We present BetterScene, an approach to enhance novel view synthesis (NVS) quality for diverse real-world scenes using extremely sparse, unconstrained photos. BetterScene leverages the production-ready Stable Video Diffusion (SVD) model pretrained on billions of frames as a strong backbone, aiming to mitigate artifacts and recover view-consistent details at inference time. Conventional methods have developed similar diffusion-based solutions to address these challenges of novel view synthesis. Despite significant improvements, these methods typically rely on off-the-shelf pretrained diffusion priors and fine-tune only the UNet module while keeping other components frozen, which still leads to inconsistent details and artifacts even when incorporating geometry-aware regularizations like depth or semantic conditions. To address this, we investigate the latent space of the diffusion model and introduce two components: (1) temporal equivariance regularization and (2) vision foundation model-aligned representation, both applied to the variational autoencoder (VAE) module within the SVD pipeline. BetterScene integrates a feed-forward 3D Gaussian Splatting (3DGS) model to render features as inputs for the SVD enhancer and generate continuous, artifact-free, consistent novel views. We evaluate on the challenging DL3DV-10K dataset and demonstrate superior performance compared to state-of-the-art methods.

Yunus Talha Erzurumlu, Jiyong Kwag, Alper Yilmaz

Cross-view geo-localization (CVGL) estimates a camera's location by matching a street-view image to geo-referenced overhead imagery, enabling GPS-denied localization and navigation. Existing methods almost universally formulate CVGL as an image-retrieval problem in a contrastively trained embedding space. This ties performance to large batches and hard negative mining, and it ignores both the geometric structure of maps and the coverage mismatch between street-view and overhead imagery. In particular, salient landmarks visible from the street view can fall outside a fixed satellite crop, making retrieval targets ambiguous and limiting explicit spatial inference over the map. We propose Just Zoom In, an alternative formulation that performs CVGL via autoregressive zooming over a city-scale overhead map. Starting from a coarse satellite view, the model takes a short sequence of zoom-in decisions to select a terminal satellite cell at a target resolution, without contrastive losses or hard negative mining. We further introduce a realistic benchmark with crowd-sourced street views and high-resolution satellite imagery that reflects real capture conditions. On this benchmark, Just Zoom In achieves state-of-the-art performance, improving Recall@1 within 50 m by 5.5% and Recall@1 within 100 m by 9.6% over the strongest contrastive-retrieval baseline. These results demonstrate the effectiveness of sequential coarse-to-fine spatial reasoning for cross-view geo-localization.

Trong-Thuan Nguyen, Pha Nguyen, Jackson Cothren et al.

Multimodal LLMs have advanced vision-language tasks but still struggle with understanding video scenes. To bridge this gap, Video Scene Graph Generation (VidSGG) has emerged to capture multi-object relationships across video frames. However, prior methods rely on pairwise connections, limiting their ability to handle complex multi-object interactions and reasoning. To this end, we propose Multimodal LLMs on a Scene HyperGraph (HyperGLM), promoting reasoning about multi-way interactions and higher-order relationships. Our approach uniquely integrates entity scene graphs, which capture spatial relationships between objects, with a procedural graph that models their causal transitions, forming a unified HyperGraph. Significantly, HyperGLM enables reasoning by injecting this unified HyperGraph into LLMs. Additionally, we introduce a new Video Scene Graph Reasoning (VSGR) dataset featuring 1.9M frames from third-person, egocentric, and drone views and supports five tasks: Scene Graph Generation, Scene Graph Anticipation, Video Question Answering, Video Captioning, and Relation Reasoning. Empirically, HyperGLM consistently outperforms state-of-the-art methods across five tasks, effectively modeling and reasoning complex relationships in diverse video scenes.

Pha Nguyen, Ngan Le, Jackson Cothren et al.

Object tracking is a fundamental task in computer vision, requiring the localization of objects of interest across video frames. Diffusion models have shown remarkable capabilities in visual generation, making them well-suited for addressing several requirements of the tracking problem. This work proposes a novel diffusion-based methodology to formulate the tracking task. Firstly, their conditional process allows for injecting indications of the target object into the generation process. Secondly, diffusion mechanics can be developed to inherently model temporal correspondences, enabling the reconstruction of actual frames in video. However, existing diffusion models rely on extensive and unnecessary mapping to a Gaussian noise domain, which can be replaced by a more efficient and stable interpolation process. Our proposed interpolation mechanism draws inspiration from classic image-processing techniques, offering a more interpretable, stable, and faster approach tailored specifically for the object tracking task. By leveraging the strengths of diffusion models while circumventing their limitations, our Diffusion-based INterpolation TrackeR (DINTR) presents a promising new paradigm and achieves a superior multiplicity on seven benchmarks across five indicator representations.

Michael Karnes, Alper Yilmaz

While deep learning has achieved remarkable success in medical imaging, the "black-box" nature of backpropagation-based models remains a significant barrier to clinical adoption. To bridge this gap, we propose Aristotelian Rapid Object Modeling (A-ROM), a framework built upon the Platonic Representation Hypothesis (PRH). This hypothesis posits that models trained on vast, diverse datasets converge toward a universal and objective representation of reality. By leveraging the generalizable metric space of pretrained Vision Transformers (ViTs), A-ROM enables the rapid modeling of novel medical concepts without the computational burden or opacity of further gradient-based fine-tuning. We replace traditional, opaque decision layers with a human-readable concept dictionary and a k-Nearest Neighbors (kNN) classifier to ensure the model's logic remains interpretable. Experiments on the MedMNIST v2 suite demonstrate that A-ROM delivers performance competitive with standard benchmarks while providing a simple and scalable, "few-shot" solution that meets the rigorous transparency demands of modern clinical environments.

Jiyong Kwag, Alper Yilmaz, Charles Toth

Road environment segmentation plays a significant role in autonomous driving. Numerous works based on Fully Convolutional Networks (FCNs) and Transformer architectures have been proposed to leverage local and global contextual learning for efficient and accurate semantic segmentation. In both architectures, the encoder often relies heavily on extracting continuous representations from the image, which limits the ability to represent meaningful discrete information. To address this limitation, we propose segmentation of the autonomous driving environment using vector quantization. Vector quantization offers three primary advantages for road environment segmentation. (1) Each continuous feature from the encoder is mapped to a discrete vector from the codebook, helping the model discover distinct features more easily than with complex continuous features. (2) Since a discrete feature acts as compressed versions of the encoder's continuous features, they also compress noise or outliers, enhancing the image segmentation task. (3) Vector quantization encourages the latent space to form coarse clusters of continuous features, forcing the model to group similar features, making the learned representations more structured for the decoding process. In this work, we combined vector quantization with the lightweight image segmentation model MobileUNETR and used it as a baseline model for comparison to demonstrate its efficiency. Through experiments, we achieved 77.0 % mIoU on Cityscapes, outperforming the baseline by 2.9 % without increasing the model's initial size or complexity.

Abdullah Turer, Yongsheng Bai, Halil Sezen et al.

Timely assessment of integrity of structures after seismic events is crucial for public safety and emergency response. This study focuses on assessing the structural damage conditions using deep learning methods to detect exposed steel reinforcement in concrete buildings and bridges after large earthquakes. Steel bars are typically exposed after concrete spalling or large flexural or shear cracks. The amount and distribution of exposed steel reinforcement is an indication of structural damage and degradation. To automatically detect exposed steel bars, new datasets of images collected after the 2023 Turkey Earthquakes were labeled to represent a wide variety of damaged concrete structures. The proposed method builds upon a deep learning framework, enhanced with fine-tuning, data augmentation, and testing on public datasets. An automated classification framework is developed that can be used to identify inside/outside buildings and structural components. Then, a YOLOv11 (You Only Look Once) model is trained to detect cracking and spalling damage and exposed bars. Another YOLO model is finetuned to distinguish different categories of structural damage levels. All these trained models are used to create a hybrid framework to automatically and reliably determine the damage levels from input images. This research demonstrates that rapid and automated damage detection following disasters is achievable across diverse damage contexts by utilizing image data collection, annotation, and deep learning approaches.

Pouyan Navard, Yasemin Ozkut, Srikar Adhikari et al.

Retinal detachment (RD) is a vision-threatening condition that requires timely intervention to preserve vision. Macular involvement -- whether the macula is still intact (macula-intact) or detached (macula-detached) -- is the key determinant of visual outcomes and treatment urgency. Point-of-care ultrasound (POCUS) offers a fast, non-invasive, cost-effective, and accessible imaging modality widely used in diverse clinical settings to detect RD. However, ultrasound image interpretation is limited by a lack of expertise among healthcare providers, especially in resource-limited settings. Deep learning offers the potential to automate ultrasound-based assessment of RD. However, there are no ML ultrasound algorithms currently available for clinical use to detect RD and no prior research has been done on assessing macular status using ultrasound in RD cases -- an essential distinction for surgical prioritization. Moreover, no public dataset currently supports macular-based RD classification using ultrasound video clips. We introduce Eye Retinal DEtachment ultraSound, ERDES, the first open-access dataset of ocular ultrasound clips labeled for (i) presence of retinal detachment and (ii) macula-intact versus macula-detached status. The dataset is intended to facilitate the development and evaluation of machine learning models for detecting retinal detachment. We also provide baseline benchmarks using multiple spatiotemporal convolutional neural network (CNN) architectures. All clips, labels, and training code are publicly available at https://osupcvlab.github.io/ERDES/.

Trong-Thuan Nguyen, Pha Nguyen, Jackson Cothren et al.

The rapid proliferation of video in applications such as autonomous driving, surveillance, and sports analytics necessitates robust methods for dynamic scene understanding. Despite advances in static scene graph generation and early attempts at video scene graph generation, previous methods often suffer from fragmented representations, failing to capture fine-grained spatial details and long-range temporal dependencies simultaneously. To address these limitations, we introduce the Temporal Hierarchical Cyclic Scene Graph (THYME) approach, which synergistically integrates hierarchical feature aggregation with cyclic temporal refinement to address these limitations. In particular, THYME effectively models multi-scale spatial context and enforces temporal consistency across frames, yielding more accurate and coherent scene graphs. In addition, we present AeroEye-v1.0, a novel aerial video dataset enriched with five types of interactivity that overcome the constraints of existing datasets and provide a comprehensive benchmark for dynamic scene graph generation. Empirically, extensive experiments on ASPIRe and AeroEye-v1.0 demonstrate that the proposed THYME approach outperforms state-of-the-art methods, offering improved scene understanding in ground-view and aerial scenarios.

Yuci Han, Charles Toth, Alper Yilmaz

The aim of this work is to develop an approach that enables Unmanned Aerial System (UAS) to efficiently learn to navigate in large-scale urban environments and transfer their acquired expertise to novel environments. To achieve this, we propose a meta-curriculum training scheme. First, meta-training allows the agent to learn a master policy to generalize across tasks. The resulting model is then fine-tuned on the downstream tasks. We organize the training curriculum in a hierarchical manner such that the agent is guided from coarse to fine towards the target task. In addition, we introduce Incremental Self-Adaptive Reinforcement learning (ISAR), an algorithm that combines the ideas of incremental learning and meta-reinforcement learning (MRL). In contrast to traditional reinforcement learning (RL), which focuses on acquiring a policy for a specific task, MRL aims to learn a policy with fast transfer ability to novel tasks. However, the MRL training process is time consuming, whereas our proposed ISAR algorithm achieves faster convergence than the conventional MRL algorithm. We evaluate the proposed methodologies in simulated environments and demonstrate that using this training philosophy in conjunction with the ISAR algorithm significantly improves the convergence speed for navigation in large-scale cities and the adaptation proficiency in novel environments.

Trong-Thuan Nguyen, Pha Nguyen, Xin Li et al.

Video scene graph generation (VidSGG) has emerged as a transformative approach to capturing and interpreting the intricate relationships among objects and their temporal dynamics in video sequences. In this paper, we introduce the new AeroEye dataset that focuses on multi-object relationship modeling in aerial videos. Our AeroEye dataset features various drone scenes and includes a visually comprehensive and precise collection of predicates that capture the intricate relationships and spatial arrangements among objects. To this end, we propose the novel Cyclic Graph Transformer (CYCLO) approach that allows the model to capture both direct and long-range temporal dependencies by continuously updating the history of interactions in a circular manner. The proposed approach also allows one to handle sequences with inherent cyclical patterns and process object relationships in the correct sequential order. Therefore, it can effectively capture periodic and overlapping relationships while minimizing information loss. The extensive experiments on the AeroEye dataset demonstrate the effectiveness of the proposed CYCLO model, demonstrating its potential to perform scene understanding on drone videos. Finally, the CYCLO method consistently achieves State-of-the-Art (SOTA) results on two in-the-wild scene graph generation benchmarks, i.e., PVSG and ASPIRe.

Michael Karnes, Alper Yilmaz

Feature extraction has always been a critical component of the computer vision field. More recently, state-of-the-art computer visions algorithms have incorporated Deep Neural Networks (DNN) in feature extracting roles, creating Deep Convolutional Activation Features (DeCAF). The transferability of DNN knowledge domains has enabled the wide use of pretrained DNN feature extraction for applications with novel object classes, especially those with limited training data. This study analyzes the general discriminability of novel object visual appearances encoded into the DeCAF space of six of the leading visual recognition DNN architectures. The results of this study characterize the Mahalanobis distances and cosine similarities between DeCAF object manifolds across two visual object tracking benchmark data sets. The backgrounds surrounding each object are also included as an object classes in the manifold analysis, providing a wider range of novel classes. This study found that different network architectures led to different network feature focuses that must to be considered in the network selection process. These results are generated from the VOT2015 and UAV123 benchmark data sets; however, the proposed methods can be applied to efficiently compare estimated network performance characteristics for any labeled visual data set.

Yongsheng Bai, Ramzi M. Abduallah, Halil Sezen et al.

This paper proposes a pipeline to automatically track and measure displacement and vibration of structural specimens during laboratory experiments. The latest Mask Regional Convolutional Neural Network (Mask R-CNN) can locate the targets and monitor their movement from videos recorded by a stationary camera. To improve precision and remove the noise, techniques such as Scale-invariant Feature Transform (SIFT) and various filters for signal processing are included. Experiments on three small-scale reinforced concrete beams and a shaking table test are utilized to verify the proposed method. Results show that the proposed deep learning method can achieve the goal to automatically and precisely measure the motion of tested structural members during laboratory experiments.

Ningli Xu, Debao Huang, Shuang Song et al.

In this paper, we present a case study that performs an unmanned aerial vehicle (UAV) based fine-scale 3D change detection and monitoring of progressive collapse performance of a building during a demolition event. Multi-temporal oblique photogrammetry images are collected with 3D point clouds generated at different stages of the demolition. The geometric accuracy of the generated point clouds has been evaluated against both airborne and terrestrial LiDAR point clouds, achieving an average distance of 12 cm and 16 cm for roof and facade respectively. We propose a hierarchical volumetric change detection framework that unifies multi-temporal UAV images for pose estimation (free of ground control points), reconstruction, and a coarse-to-fine 3D density change analysis. This work has provided a solution capable of addressing change detection on full 3D time-series datasets where dramatic scene content changes are presented progressively. Our change detection results on the building demolition event have been evaluated against the manually marked ground-truth changes and have achieved an F-1 score varying from 0.78 to 0.92, with consistently high precision (0.92 - 0.99). Volumetric changes through the demolition progress are derived from change detection and have shown to favorably reflect the qualitative and quantitative building demolition progression.

Shehan Perera, Srikar Adhikari, Alper Yilmaz

The rapid and seemingly endless expansion of COVID-19 can be traced back to the inefficiency and shortage of testing kits that offer accurate results in a timely manner. An emerging popular technique, which adopts improvements made in mobile ultrasound technology, allows for healthcare professionals to conduct rapid screenings on a large scale. We present an image-based solution that aims at automating the testing process which allows for rapid mass testing to be conducted with or without a trained medical professional that can be applied to rural environments and third world countries. Our contributions towards rapid large-scale testing include a novel deep learning architecture capable of analyzing ultrasound data that can run in real-time and significantly improve the current state-of-the-art detection accuracies using image-based COVID-19 detection.

Bing Zha, Alessandro Vanni, Yassin Hassan et al.

A challenging part of dynamic probabilistic risk assessment for nuclear power plants is the need for large amounts of temporal simulations given various initiating events and branching conditions from which representative feature extraction becomes complicated for subsequent applications. Artificial Intelligence techniques have been shown to be powerful tools in time-dependent sequential data processing to automatically extract and yield complex features from large data. An advanced temporal neural network referred to as the Transformer is used within a supervised learning fashion to model the time-dependent NPP simulation data and to infer whether a given sequence of events leads to core damage or not. The training and testing datasets for the Transformer are obtained by running 10,000 RELAP5-3D NPP blackout simulations with the list of variables obtained from the RAVEN software. Each simulation is classified as "OK" or "CORE DAMAGE" based on the consequence. The results show that the Transformer can learn the characteristics of the sequential data and yield promising performance with approximately 99% classification accuracy on the testing dataset.

Yongsheng Bai, Halil Sezen, Alper Yilmaz

Robust Mask R-CNN (Mask Regional Convolu-tional Neural Network) methods are proposed and tested for automatic detection of cracks on structures or their components that may be damaged during extreme events, such as earth-quakes. We curated a new dataset with 2,021 labeled images for training and validation and aimed to find end-to-end deep neural networks for crack detection in the field. With data augmentation and parameters fine-tuning, Path Aggregation Network (PANet) with spatial attention mechanisms and High-resolution Network (HRNet) are introduced into Mask R-CNNs. The tests on three public datasets with low- or high-resolution images demonstrate that the proposed methods can achieve a big improvement over alternative networks, so the proposed method may be sufficient for crack detection for a variety of scales in real applications.

Bing Zha, Alper Yilmaz

In this paper, we propose a novel trajectory learning method that exploits motion trajectories on topological map using recurrent neural network for temporally consistent geolocalization of object. Inspired by human's ability to both be aware of distance and direction of self-motion in navigation, our trajectory learning method learns a pattern representation of trajectories encoded as a sequence of distances and turning angles to assist self-localization. We pose the learning process as a conditional sequence prediction problem in which each output locates the object on a traversable path in a map. Considering the prediction sequence ought to be topologically connected in the graph-structured map, we adopt two different hypotheses generation and elimination strategies to eliminate disconnected sequence prediction. We demonstrate our approach on the KITTI stereo visual odometry dataset which is a city-scale environment and can generate trajectory with metric information. The key benefits of our approach to geolocalization are that 1) we take advantage of powerful sequence modeling ability of recurrent neural network and its robustness to noisy input, 2) only require a map in the form of a graph and simply use an affordable sensor that generates motion trajectory and 3) do not need initial position. The experiments show that the motion trajectories can be learned by training an recurrent neural network, and temporally consistent geolocation can be predicted with both of the proposed strategies.

Ye Lyu, George Vosselman, Guisong Xia et al.

Semantic segmentation has been one of the leading research interests in computer vision recently. It serves as a perception foundation for many fields, such as robotics and autonomous driving. The fast development of semantic segmentation attributes enormously to the large scale datasets, especially for the deep learning related methods. There already exist several semantic segmentation datasets for comparison among semantic segmentation methods in complex urban scenes, such as the Cityscapes and CamVid datasets, where the side views of the objects are captured with a camera mounted on the driving car. There also exist semantic labeling datasets for the airborne images and the satellite images, where the top views of the objects are captured. However, only a few datasets capture urban scenes from an oblique Unmanned Aerial Vehicle (UAV) perspective, where both of the top view and the side view of the objects can be observed, providing more information for object recognition. In this paper, we introduce our UAVid dataset, a new high-resolution UAV semantic segmentation dataset as a complement, which brings new challenges, including large scale variation, moving object recognition and temporal consistency preservation. Our UAV dataset consists of 30 video sequences capturing 4K high-resolution images in slanted views. In total, 300 images have been densely labeled with 8 classes for the semantic labeling task. We have provided several deep learning baseline methods with pre-training, among which the proposed Multi-Scale-Dilation net performs the best via multi-scale feature extraction. Our UAVid website and the labeling tool have been published https://uavid.nl/.

Oliver Nina, Washington Garcia, Scott Clouse et al.

Learning visual feature representations for video analysis is a daunting task that requires a large amount of training samples and a proper generalization framework. Many of the current state of the art methods for video captioning and movie description rely on simple encoding mechanisms through recurrent neural networks to encode temporal visual information extracted from video data. In this paper, we introduce a novel multitask encoder-decoder framework for automatic semantic description and captioning of video sequences. In contrast to current approaches, our method relies on distinct decoders that train a visual encoder in a multitask fashion. Our system does not depend solely on multiple labels and allows for a lack of training data working even with datasets where only one single annotation is viable per video. Our method shows improved performance over current state of the art methods in several metrics on multi-caption and single-caption datasets. To the best of our knowledge, our method is the first method to use a multitask approach for encoding video features. Our method demonstrates its robustness on the Large Scale Movie Description Challenge (LSMDC) 2017 where our method won the movie description task and its results were ranked among other competitors as the most helpful for the visually impaired.

Shirui Li, Alper Yilmaz, Changlin Xiao et al.

In this paper, we propose a new method to detect 4D spatiotemporal interest points though an implicit surface, we refer to as the 4D-ISIP. We use a 3D volume which has a truncated signed distance function(TSDF) for every voxel to represent our 3D object model. The TSDF represents the distance between the spatial points and object surface points which is an implicit surface representation. Our novelty is to detect the points where the local neighborhood has significant variations along both spatial and temporal directions. We established a system to acquire 3D human motion dataset using only one Kinect. Experimental results show that our method can detect 4D-ISIP for different human actions.