Christian Micheloni

Matteo Dunnhofer, Antonino Furnari, Giovanni Maria Farinella et al.

The understanding of human-object interactions is fundamental in First Person Vision (FPV). Visual tracking algorithms which follow the objects manipulated by the camera wearer can provide useful information to effectively model such interactions. In the last years, the computer vision community has significantly improved the performance of tracking algorithms for a large variety of target objects and scenarios. Despite a few previous attempts to exploit trackers in the FPV domain, a methodical analysis of the performance of state-of-the-art trackers is still missing. This research gap raises the question of whether current solutions can be used ``off-the-shelf'' or more domain-specific investigations should be carried out. This paper aims to provide answers to such questions. We present the first systematic investigation of single object tracking in FPV. Our study extensively analyses the performance of 42 algorithms including generic object trackers and baseline FPV-specific trackers. The analysis is carried out by focusing on different aspects of the FPV setting, introducing new performance measures, and in relation to FPV-specific tasks. The study is made possible through the introduction of TREK-150, a novel benchmark dataset composed of 150 densely annotated video sequences. Our results show that object tracking in FPV poses new challenges to current visual trackers. We highlight the factors causing such behavior and point out possible research directions. Despite their difficulties, we prove that trackers bring benefits to FPV downstream tasks requiring short-term object tracking. We expect that generic object tracking will gain popularity in FPV as new and FPV-specific methodologies are investigated.

Moritz Nottebaum, Matteo Dunnhofer, Christian Micheloni

Vision backbone networks play a central role in modern computer vision. Enhancing their efficiency directly benefits a wide range of downstream applications. To measure efficiency, many publications rely on MACs (Multiply Accumulate operations) as a predictor of execution time. In this paper, we experimentally demonstrate the shortcomings of such a metric, especially in the context of edge devices. By contrasting the MAC count and execution time of common architectural design elements, we identify key factors for efficient execution and provide insights to optimize backbone design. Based on these insights, we present LowFormer, a novel vision backbone family. LowFormer features a streamlined macro and micro design that includes Lowtention, a lightweight alternative to Multi-Head Self-Attention. Lowtention not only proves more efficient, but also enables superior results on ImageNet. Additionally, we present an edge GPU version of LowFormer, that can further improve upon its baseline's speed on edge GPU and desktop GPU. We demonstrate LowFormer's wide applicability by evaluating it on smaller image classification datasets, as well as adapting it to several downstream tasks, such as object detection, semantic segmentation, image retrieval, and visual object tracking. LowFormer models consistently achieve remarkable speed-ups across various hardware platforms compared to recent state-of-the-art backbones. Code and models are available at https://github.com/altair199797/LowFormer/blob/main/Beyond_MACs.md.

Matteo Dunnhofer, Christian Micheloni

How to combine the complementary capabilities of an ensemble of different algorithms has been of central interest in visual object tracking. A significant progress on such a problem has been achieved, but considering short-term tracking scenarios. Instead, long-term tracking settings have been substantially ignored by the solutions. In this paper, we explicitly consider long-term tracking scenarios and provide a framework, named CoCoLoT, that combines the characteristics of complementary visual trackers to achieve enhanced long-term tracking performance. CoCoLoT perceives whether the trackers are following the target object through an online learned deep verification model, and accordingly activates a decision policy which selects the best performing tracker as well as it corrects the performance of the failing one. The proposed methodology is evaluated extensively and the comparison with several other solutions reveals that it competes favourably with the state-of-the-art on the most popular long-term visual tracking benchmarks.

Matteo Dunnhofer, Luca Sordi, Christian Micheloni

Trajectories are fundamental to winning in alpine skiing. Tools enabling the analysis of such curves can enhance the training activity and enrich broadcasting content. In this paper, we propose SkiTraVis, an algorithm to visualize the sequence of points traversed by a skier during its performance. SkiTraVis works on monocular videos and constitutes a pipeline of a visual tracker to model the skier's motion and of a frame correspondence module to estimate the camera's motion. The separation of the two motions enables the visualization of the trajectory according to the moving camera's perspective. We performed experiments on videos of real-world professional competitions to quantify the visualization error, the computational efficiency, as well as the applicability. Overall, the results achieved demonstrate the potential of our solution for broadcasting media enhancement and coach assistance.

Moritz Nottebaum, Matteo Dunnhofer, Christian Micheloni

Recent research on vision backbone architectures has predominantly focused on optimizing efficiency for hardware platforms with high parallel processing capabilities. This category increasingly includes embedded systems such as mobile phones and embedded AI accelerator modules. In contrast, CPUs do not have the possibility to parallelize operations in the same manner, wherefore models benefit from a specific design philosophy that balances amount of operations (MACs) and hardware-efficient execution by having high MACs per second (MACpS). In pursuit of this, we investigate two modifications to standard convolutions, aimed at reducing computational cost: grouping convolutions and reducing kernel sizes. While both adaptations substantially decrease the total number of MACs required for inference, sustaining low latency necessitates preserving hardware-efficiency. Our experiments across diverse CPU devices confirm that these adaptations successfully retain high hardware-efficiency on CPUs. Based on these insights, we introduce CPUBone, a new family of vision backbone models optimized for CPU-based inference. CPUBone achieves state-of-the-art Speed-Accuracy Trade-offs (SATs) across a wide range of CPU devices and effectively transfers its efficiency to downstream tasks such as object detection and semantic segmentation. Models and code are available at https://github.com/altair199797/CPUBone.

Moritz Nottebaum, Matteo Dunnhofer, Christian Micheloni

Research in efficient vision backbones is evolving into models that are a mixture of convolutions and transformer blocks. A smart combination of both, architecture-wise and component-wise is mandatory to excel in the speedaccuracy trade-off. Most publications focus on maximizing accuracy and utilize MACs (multiply accumulate operations) as an efficiency metric. The latter however often do not measure accurately how fast a model actually is due to factors like memory access cost and degree of parallelism. We analyzed common modules and architectural design choices for backbones not in terms of MACs, but rather in actual throughput and latency, as the combination of the latter two is a better representation of the efficiency of models in real applications. We applied the conclusions taken from that analysis to create a recipe for increasing hardware-efficiency in macro design. Additionally we introduce a simple slimmed-down version of MultiHead Self-Attention, that aligns with our analysis. We combine both macro and micro design to create a new family of hardware-efficient backbone networks called LowFormer. LowFormer achieves a remarkable speedup in terms of throughput and latency, while achieving similar or better accuracy than current state-of-the-art efficient backbones. In order to prove the generalizability of our hardware-efficient design, we evaluate our method on GPU, mobile GPU and ARM CPU. We further show that the downstream tasks object detection and semantic segmentation profit from our hardware-efficient architecture. Code and models are available at https://github.com/ altair199797/LowFormer.

Belén Serrano-Antón, Mubashara Rehman, Niki Martinel et al.

For the planning of radiotherapy treatments for head and neck cancers, Computed Tomography (CT) scans of the patients are typically employed. However, in patients with head and neck cancer, the quality of standard CT scans generated using kilo-Voltage (kVCT) tube potentials is severely degraded by streak artifacts occurring in the presence of metallic implants such as dental fillings. Some radiotherapy devices offer the possibility of acquiring Mega-Voltage CT (MVCT) for daily patient setup verification, due to the higher energy of X-rays used, MVCT scans are almost entirely free from artifacts making them more suitable for radiotherapy treatment planning. In this study, we leverage the advantages of kVCT scans with those of MVCT scans (artifact-free). We propose a deep learning-based approach capable of generating artifact-free MVCT images from acquired kVCT images. The outcome offers the benefits of artifact-free MVCT images with enhanced soft tissue contrast, harnessing valuable information obtained through kVCT technology for precise therapy calibration. Our proposed method employs UNet-inspired model, and is compared with adversarial learning and transformer networks. This first and unique approach achieves remarkable success, with PSNR of 30.02 dB across the entire patient volume and 27.47 dB in artifact-affected regions exclusively. It is worth noting that the PSNR calculation excludes the background, concentrating solely on the region of interest.

Rao Muhammad Umer, Christian Micheloni

The current existing deep image super-resolution methods usually assume that a Low Resolution (LR) image is bicubicly downscaled of a High Resolution (HR) image. However, such an ideal bicubic downsampling process is different from the real LR degradations, which usually come from complicated combinations of different degradation processes, such as camera blur, sensor noise, sharpening artifacts, JPEG compression, and further image editing, and several times image transmission over the internet and unpredictable noises. It leads to the highly ill-posed nature of the inverse upscaling problem. To address these issues, we propose a GAN-based SR approach with learnable adaptive sinusoidal nonlinearities incorporated in LR and SR models by directly learn degradation distributions and then synthesize paired LR/HR training data to train the generalized SR model to real image degradations. We demonstrate the effectiveness of our proposed approach in quantitative and qualitative experiments.

Rita Pucci, Christian Micheloni, Niki Martinel

The degradation in the underwater images is due to wavelength-dependent light attenuation, scattering, and to the diversity of the water types in which they are captured. Deep neural networks take a step in this field, providing autonomous models able to achieve the enhancement of underwater images. We introduce Underwater Capsules Vectors GAN UWCVGAN based on the discrete features quantization paradigm from VQGAN for this task. The proposed UWCVGAN combines an encoding network, which compresses the image into its latent representation, with a decoding network, able to reconstruct the enhancement of the image from the only latent representation. In contrast with VQGAN, UWCVGAN achieves feature quantization by exploiting the clusterization ability of capsule layer, making the model completely trainable and easier to manage. The model obtains enhanced underwater images with high quality and fine details. Moreover, the trained encoder is independent of the decoder giving the possibility to be embedded onto the collector as compressing algorithm to reduce the memory space required for the images, of factor $3\times$. \myUWCVGAN{ }is validated with quantitative and qualitative analysis on benchmark datasets, and we present metrics results compared with the state of the art.

Mubashara Rehman, Niki Martinel, Michele Avanzo et al.

Metal artifacts in computed tomography (CT) severely degrade image quality, compromising diagnostic accuracy and radiotherapy planning, especially in cancer patients with high-density implants. We propose H3D-MarNet, a two-stage framework for artifact-aware CT domain transformation from kilo-voltage CT (kVCT) to mega-voltage CT (MVCT). In the first stage, a wavelet-based preprocessing module suppresses metal-induced artifacts through frequency-aware denoising while preserving anatomical structures. In second stage, Domain-TransNet performs kVCT-to-MVCT domain transformation using a hybrid volumetric learning architecture. Domain-TransNet integrates a CNN-based encoder to capture fine-grained local anatomical details and a transformer-based encoder to model long-range volumetric dependencies. The complementary representations are fused through an attention-based feature fusion mechanism to ensure spatial and contextual coherence across slices. A multi-stage, attention-guided decoder, supported by deep supervision, progressively reconstructs artifact-suppressed MVCT volumes. Extensive experiments demonstrate that H3D-MarNet achieves 28.14 dB PSNR and 0.717 SSIM on artifact-affected slices from full dataset, indicating effective metal artifact suppression and anatomical preservation, highlighting its potential for reliable CT modality transformation in clinical radiotherapy workflows.

Zheng Chen, Zongwei Wu, Eduard Zamfir et al.

This paper reviews the NTIRE 2024 challenge on image super-resolution ($\times$4), highlighting the solutions proposed and the outcomes obtained. The challenge involves generating corresponding high-resolution (HR) images, magnified by a factor of four, from low-resolution (LR) inputs using prior information. The LR images originate from bicubic downsampling degradation. The aim of the challenge is to obtain designs/solutions with the most advanced SR performance, with no constraints on computational resources (e.g., model size and FLOPs) or training data. The track of this challenge assesses performance with the PSNR metric on the DIV2K testing dataset. The competition attracted 199 registrants, with 20 teams submitting valid entries. This collective endeavour not only pushes the boundaries of performance in single-image SR but also offers a comprehensive overview of current trends in this field.

Matteo Dunnhofer, Luca Sordi, Niki Martinel et al.

Skiing is a popular winter sport discipline with a long history of competitive events. In this domain, computer vision has the potential to enhance the understanding of athletes' performance, but its application lags behind other sports due to limited studies and datasets. This paper makes a step forward in filling such gaps. A thorough investigation is performed on the task of skier tracking in a video capturing his/her complete performance. Obtaining continuous and accurate skier localization is preemptive for further higher-level performance analyses. To enable the study, the largest and most annotated dataset for computer vision in skiing, SkiTB, is introduced. Several visual object tracking algorithms, including both established methodologies and a newly introduced skier-optimized baseline algorithm, are tested using the dataset. The results provide valuable insights into the applicability of different tracking methods for vision-based skiing analysis. SkiTB, code, and results are available at https://machinelearning.uniud.it/datasets/skitb.

Zaira Manigrasso, Matteo Dunnhofer, Antonino Furnari et al.

Episodic memory retrieval enables wearable cameras to recall objects or events previously observed in video. However, existing formulations assume an "offline" setting with full video access at query time, limiting their applicability in real-world scenarios with power and storage-constrained wearable devices. Towards more application-ready episodic memory systems, we introduce Online Visual Query 2D (OVQ2D), a task where models process video streams online, observing each frame only once, and retrieve object localizations using a compact memory instead of full video history. We address OVQ2D with ESOM (Egocentric Streaming Object Memory), a novel framework integrating an object discovery module, an object tracking module, and a memory module that find, track, and store spatio-temporal object information for efficient querying. Experiments on Ego4D demonstrate ESOM's superiority over other online approaches, though OVQ2D remains challenging, with top performance at only ~4% success. ESOM's accuracy increases markedly with perfect object tracking (31.91%), discovery (40.55%), or both (81.92%), underscoring the need of applied research on these components.

Niki Martinel, Mariano Serrao, Christian Micheloni

We introduce a novel state-space model (SSM)-based framework for skeleton-based human action recognition, with an anatomically-guided architecture that improves state-of-the-art performance in both clinical diagnostics and general action recognition tasks. Our approach decomposes skeletal motion analysis into spatial, temporal, and spatio-temporal streams, using channel partitioning to capture distinct movement characteristics efficiently. By implementing a structured, multi-directional scanning strategy within SSMs, our model captures local joint interactions and global motion patterns across multiple anatomical body parts. This anatomically-aware decomposition enhances the ability to identify subtle motion patterns critical in medical diagnosis, such as gait anomalies associated with neurological conditions. On public action recognition benchmarks, i.e., NTU RGB+D, NTU RGB+D 120, and NW-UCLA, our model outperforms current state-of-the-art methods, achieving accuracy improvements up to $3.2\%$ with lower computational complexity than previous leading transformer-based models. We also introduce a novel medical dataset for motion-based patient neurological disorder analysis to validate our method's potential in automated disease diagnosis.

Matteo Dunnhofer, Zaira Manigrasso, Christian Micheloni

Visual object tracking and segmentation are becoming fundamental tasks for understanding human activities in egocentric vision. Recent research has benchmarked state-of-the-art methods and concluded that first person egocentric vision presents challenges compared to previously studied domains. However, these claims are based on evaluations conducted across significantly different scenarios. Many of the challenging characteristics attributed to egocentric vision are also present in third person videos of human-object activities. This raises a critical question: how much of the observed performance drop stems from the unique first person viewpoint inherent to egocentric vision versus the domain of human-object activities? To address this question, we introduce a new benchmark study designed to disentangle such factors. Our evaluation strategy enables a more precise separation of challenges related to the first person perspective from those linked to the broader domain of human-object activity understanding. By doing so, we provide deeper insights into the true sources of difficulty in egocentric tracking and segmentation, facilitating more targeted advancements on this task.

Mubashara Rehman, Niki Martinel, Michele Avanzo et al.

Artifacts in kilo-Voltage CT (kVCT) imaging degrade image quality, impacting clinical decisions. We propose a deep learning framework for metal artifact reduction (MAR) and domain transformation from kVCT to Mega-Voltage CT (MVCT). The proposed framework, ReMAR-DS, utilizes an encoder-decoder architecture with enhanced feature recalibration, effectively reducing artifacts while preserving anatomical structures. This ensures that only relevant information is utilized in the reconstruction process. By infusing recalibrated features from the encoder block, the model focuses on relevant spatial regions (e.g., areas with artifacts) and highlights key features across channels (e.g., anatomical structures), leading to improved reconstruction of artifact-corrupted regions. Unlike traditional MAR methods, our approach bridges the gap between high-resolution kVCT and artifact-resistant MVCT, enhancing radiotherapy planning. It produces high-quality MVCT-like reconstructions, validated through qualitative and quantitative evaluations. Clinically, this enables oncologists to rely on kVCT alone, reducing repeated high-dose MVCT scans and lowering radiation exposure for cancer patients.

Matteo Dunnhofer, Alberto Zurini, Maurizio Dunnhofer et al.

Trajectories are fundamental in different skiing disciplines. Tools enabling the analysis of such curves can enhance the training activity and enrich the broadcasting contents. However, the solutions currently available are based on geo-localized sensors and surface models. In this short paper, we propose a video-based approach to reconstruct the sequence of points traversed by an athlete during its performance. Our prototype is constituted by a pipeline of deep learning-based algorithms to reconstruct the athlete's motion and to visualize it according to the camera perspective. This is achieved for different skiing disciplines in the wild without any camera calibration. We tested our solution on broadcast and smartphone-captured videos of alpine skiing and ski jumping professional competitions. The qualitative results achieved show the potential of our solution.

Rao Muhammad Umer, Christian Micheloni

Modern digital cameras and smartphones mostly rely on image signal processing (ISP) pipelines to produce realistic colored RGB images. However, compared to DSLR cameras, low-quality images are usually obtained in many portable mobile devices with compact camera sensors due to their physical limitations. The low-quality images have multiple degradations i.e., sub-pixel shift due to camera motion, mosaick patterns due to camera color filter array, low-resolution due to smaller camera sensors, and the rest information are corrupted by the noise. Such degradations limit the performance of current Single Image Super-resolution (SISR) methods in recovering high-resolution (HR) image details from a single low-resolution (LR) image. In this work, we propose a Raw Burst Super-Resolution Iterative Convolutional Neural Network (RBSRICNN) that follows the burst photography pipeline as a whole by a forward (physical) model. The proposed Burst SR scheme solves the problem with classical image regularization, convex optimization, and deep learning techniques, compared to existing black-box data-driven methods. The proposed network produces the final output by an iterative refinement of the intermediate SR estimates. We demonstrate the effectiveness of our proposed approach in quantitative and qualitative experiments that generalize robustly to real LR burst inputs with onl synthetic burst data available for training.

Asad Munir, Chengjin Lyu, Bart Goossens et al.

Person re-identification (re-id) aims to retrieve images of same identities across different camera views. Resolution mismatch occurs due to varying distances between person of interest and cameras, this significantly degrades the performance of re-id in real world scenarios. Most of the existing approaches resolve the re-id task as low resolution problem in which a low resolution query image is searched in a high resolution images gallery. Several approaches apply image super resolution techniques to produce high resolution images but ignore the multiple resolutions of gallery images which is a better realistic scenario. In this paper, we introduce channel correlations to improve the learning of features from the degraded data. In addition, to overcome the problem of multiple resolutions we propose a Resolution based Feature Distillation (RFD) approach. Such an approach learns resolution invariant features by filtering the resolution related features from the final feature vectors that are used to compute the distance matrix. We tested the proposed approach on two synthetically created datasets and on one original multi resolution dataset with real degradation. Our approach improves the performance when multiple resolutions occur in the gallery and have comparable results in case of single resolution (low resolution re-id).

Matteo Dunnhofer, Antonino Furnari, Giovanni Maria Farinella et al.

Understanding human-object interactions is fundamental in First Person Vision (FPV). Tracking algorithms which follow the objects manipulated by the camera wearer can provide useful cues to effectively model such interactions. Visual tracking solutions available in the computer vision literature have significantly improved their performance in the last years for a large variety of target objects and tracking scenarios. However, despite a few previous attempts to exploit trackers in FPV applications, a methodical analysis of the performance of state-of-the-art trackers in this domain is still missing. In this paper, we fill the gap by presenting the first systematic study of object tracking in FPV. Our study extensively analyses the performance of recent visual trackers and baseline FPV trackers with respect to different aspects and considering a new performance measure. This is achieved through TREK-150, a novel benchmark dataset composed of 150 densely annotated video sequences. Our results show that object tracking in FPV is challenging, which suggests that more research efforts should be devoted to this problem so that tracking could benefit FPV tasks.

Rao Muhammad Umer, Asad Munir, Christian Micheloni

Recently, most of state-of-the-art single image super-resolution (SISR) methods have attained impressive performance by using deep convolutional neural networks (DCNNs). The existing SR methods have limited performance due to a fixed degradation settings, i.e. usually a bicubic downscaling of low-resolution (LR) image. However, in real-world settings, the LR degradation process is unknown which can be bicubic LR, bilinear LR, nearest-neighbor LR, or real LR. Therefore, most SR methods are ineffective and inefficient in handling more than one degradation settings within a single network. To handle the multiple degradation, i.e. refers to multi-domain image super-resolution, we propose a deep Super-Resolution Residual StarGAN (SR2*GAN), a novel and scalable approach that super-resolves the LR images for the multiple LR domains using only a single model. The proposed scheme is trained in a StarGAN like network topology with a single generator and discriminator networks. We demonstrate the effectiveness of our proposed approach in quantitative and qualitative experiments compared to other state-of-the-art methods.

Goutam Bhat, Martin Danelljan, Radu Timofte et al.

This paper reviews the NTIRE2021 challenge on burst super-resolution. Given a RAW noisy burst as input, the task in the challenge was to generate a clean RGB image with 4 times higher resolution. The challenge contained two tracks; Track 1 evaluating on synthetically generated data, and Track 2 using real-world bursts from mobile camera. In the final testing phase, 6 teams submitted results using a diverse set of solutions. The top-performing methods set a new state-of-the-art for the burst super-resolution task.

Matteo Dunnhofer, Niki Martinel, Christian Micheloni

Deep regression trackers are among the fastest tracking algorithms available, and therefore suitable for real-time robotic applications. However, their accuracy is inadequate in many domains due to distribution shift and overfitting. In this paper we overcome such limitations by presenting the first methodology for domain adaption of such a class of trackers. To reduce the labeling effort we propose a weakly-supervised adaptation strategy, in which reinforcement learning is used to express weak supervision as a scalar application-dependent and temporally-delayed feedback. At the same time, knowledge distillation is employed to guarantee learning stability and to compress and transfer knowledge from more powerful but slower trackers. Extensive experiments on five different robotic vision domains demonstrate the relevance of our methodology. Real-time speed is achieved on embedded devices and on machines without GPUs, while accuracy reaches significant results.

Rita Pucci, Christian Micheloni, Niki Martinel

Image colourisation is an ill-posed problem, with multiple correct solutions which depend on the context and object instances present in the input datum. Previous approaches attacked the problem either by requiring intense user interactions or by exploiting the ability of convolutional neural networks (CNNs) in learning image level (context) features. However, obtaining human hints is not always feasible and CNNs alone are not able to learn object-level semantics unless multiple models pretrained with supervision are considered. In this work, we propose a single network, named UCapsNet, that separate image-level features obtained through convolutions and object-level features captured by means of capsules. Then, by skip connections over different layers, we enforce collaboration between such disentangling factors to produce high quality and plausible image colourisation. We pose the problem as a classification task that can be addressed by a fully self-supervised approach, thus requires no human effort. Experimental results on three benchmark datasets show that our approach outperforms existing methods on standard quality metrics and achieves a state of the art performances on image colourisation. A large scale user study shows that our method is preferred over existing solutions.

Rita Pucci, Christian Micheloni, Gian Luca Foresti et al.

Human beings can imagine the colours of a grayscale image with no particular effort thanks to their ability of semantic feature extraction. Can an autonomous system achieve that? Can it hallucinate plausible and vibrant colours? This is the colourisation problem. Different from existing works relying on convolutional neural network models pre-trained with supervision, we cast such colourisation problem as a self-supervised learning task. We tackle the problem with the introduction of a novel architecture based on Capsules trained following the adversarial learning paradigm. Capsule networks are able to extract a semantic representation of the entities in the image but loose details about their spatial information, which is important for colourising a grayscale image. Thus our UCapsNet structure comes with an encoding phase that extracts entities through capsules and spatial details through convolutional neural networks. A decoding phase merges the entity features with the spatial features to hallucinate a plausible colour version of the input datum. Results on the ImageNet benchmark show that our approach is able to generate more vibrant and plausible colours than exiting solutions and achieves superior performance than models pre-trained with supervision.

Matteo Dunnhofer, Antonino Furnari, Giovanni Maria Farinella et al.

Understanding human-object interactions is fundamental in First Person Vision (FPV). Tracking algorithms which follow the objects manipulated by the camera wearer can provide useful cues to effectively model such interactions. Despite a few previous attempts to exploit trackers in FPV applications, a methodical analysis of the performance of state-of-the-art visual trackers in this domain is still missing. In this short paper, we provide a recap of the first systematic study of object tracking in FPV. Our work extensively analyses the performance of recent and baseline FPV trackers with respect to different aspects. This is achieved through TREK-150, a novel benchmark dataset composed of 150 densely annotated video sequences. The results suggest that more research efforts should be devoted to this problem so that tracking could benefit FPV tasks. The full version of this paper is available at arXiv:2108.13665.

Pengxu Wei, Hannan Lu, Radu Timofte et al.

This paper introduces the real image Super-Resolution (SR) challenge that was part of the Advances in Image Manipulation (AIM) workshop, held in conjunction with ECCV 2020. This challenge involves three tracks to super-resolve an input image for $\times$2, $\times$3 and $\times$4 scaling factors, respectively. The goal is to attract more attention to realistic image degradation for the SR task, which is much more complicated and challenging, and contributes to real-world image super-resolution applications. 452 participants were registered for three tracks in total, and 24 teams submitted their results. They gauge the state-of-the-art approaches for real image SR in terms of PSNR and SSIM.

Kai Zhang, Martin Danelljan, Yawei Li et al.

This paper reviews the AIM 2020 challenge on efficient single image super-resolution with focus on the proposed solutions and results. The challenge task was to super-resolve an input image with a magnification factor x4 based on a set of prior examples of low and corresponding high resolution images. The goal is to devise a network that reduces one or several aspects such as runtime, parameter count, FLOPs, activations, and memory consumption while at least maintaining PSNR of MSRResNet. The track had 150 registered participants, and 25 teams submitted the final results. They gauge the state-of-the-art in efficient single image super-resolution.

Rao Muhammad Umer, Gian Luca Foresti, Christian Micheloni

Deep convolutional neural networks (CNNs) have recently achieved great success for single image super-resolution (SISR) task due to their powerful feature representation capabilities. The most recent deep learning based SISR methods focus on designing deeper / wider models to learn the non-linear mapping between low-resolution (LR) inputs and high-resolution (HR) outputs. These existing SR methods do not take into account the image observation (physical) model and thus require a large number of network's trainable parameters with a great volume of training data. To address these issues, we propose a deep Iterative Super-Resolution Residual Convolutional Network (ISRResCNet) that exploits the powerful image regularization and large-scale optimization techniques by training the deep network in an iterative manner with a residual learning approach. Extensive experimental results on various super-resolution benchmarks demonstrate that our method with a few trainable parameters improves the results for different scaling factors in comparison with the state-of-art methods.

Rao Muhammad Umer, Christian Micheloni

Recent deep learning based single image super-resolution (SISR) methods mostly train their models in a clean data domain where the low-resolution (LR) and the high-resolution (HR) images come from noise-free settings (same domain) due to the bicubic down-sampling assumption. However, such degradation process is not available in real-world settings. We consider a deep cyclic network structure to maintain the domain consistency between the LR and HR data distributions, which is inspired by the recent success of CycleGAN in the image-to-image translation applications. We propose the Super-Resolution Residual Cyclic Generative Adversarial Network (SRResCycGAN) by training with a generative adversarial network (GAN) framework for the LR to HR domain translation in an end-to-end manner. We demonstrate our proposed approach in the quantitative and qualitative experiments that generalize well to the real image super-resolution and it is easy to deploy for the mobile/embedded devices. In addition, our SR results on the AIM 2020 Real Image SR Challenge datasets demonstrate that the proposed SR approach achieves comparable results as the other state-of-art methods.

Matteo Dunnhofer, Niki Martinel, Christian Micheloni

Visual object tracking is the problem of predicting a target object's state in a video. Generally, bounding-boxes have been used to represent states, and a surge of effort has been spent by the community to produce efficient causal algorithms capable of locating targets with such representations. As the field is moving towards binary segmentation masks to define objects more precisely, in this paper we propose to extensively explore target-conditioned segmentation methods available in the computer vision community, in order to transform any bounding-box tracker into a segmentation tracker. Our analysis shows that such methods allow trackers to compete with recently proposed segmentation trackers, while performing quasi real-time.

Matteo Dunnhofer, Niki Martinel, Christian Micheloni

Visual object tracking was generally tackled by reasoning independently on fast processing algorithms, accurate online adaptation methods, and fusion of trackers. In this paper, we unify such goals by proposing a novel tracking methodology that takes advantage of other visual trackers, offline and online. A compact student model is trained via the marriage of knowledge distillation and reinforcement learning. The first allows to transfer and compress tracking knowledge of other trackers. The second enables the learning of evaluation measures which are then exploited online. After learning, the student can be ultimately used to build (i) a very fast single-shot tracker, (ii) a tracker with a simple and effective online adaptation mechanism, (iii) a tracker that performs fusion of other trackers. Extensive validation shows that the proposed algorithms compete with real-time state-of-the-art trackers.

Andreas Lugmayr, Martin Danelljan, Radu Timofte et al.

This paper reviews the NTIRE 2020 challenge on real world super-resolution. It focuses on the participating methods and final results. The challenge addresses the real world setting, where paired true high and low-resolution images are unavailable. For training, only one set of source input images is therefore provided along with a set of unpaired high-quality target images. In Track 1: Image Processing artifacts, the aim is to super-resolve images with synthetically generated image processing artifacts. This allows for quantitative benchmarking of the approaches \wrt a ground-truth image. In Track 2: Smartphone Images, real low-quality smart phone images have to be super-resolved. In both tracks, the ultimate goal is to achieve the best perceptual quality, evaluated using a human study. This is the second challenge on the subject, following AIM 2019, targeting to advance the state-of-the-art in super-resolution. To measure the performance we use the benchmark protocol from AIM 2019. In total 22 teams competed in the final testing phase, demonstrating new and innovative solutions to the problem.

Rao Muhammad Umer, Gian Luca Foresti, Christian Micheloni

Most current deep learning based single image super-resolution (SISR) methods focus on designing deeper / wider models to learn the non-linear mapping between low-resolution (LR) inputs and the high-resolution (HR) outputs from a large number of paired (LR/HR) training data. They usually take as assumption that the LR image is a bicubic down-sampled version of the HR image. However, such degradation process is not available in real-world settings i.e. inherent sensor noise, stochastic noise, compression artifacts, possible mismatch between image degradation process and camera device. It reduces significantly the performance of current SISR methods due to real-world image corruptions. To address these problems, we propose a deep Super-Resolution Residual Convolutional Generative Adversarial Network (SRResCGAN) to follow the real-world degradation settings by adversarial training the model with pixel-wise supervision in the HR domain from its generated LR counterpart. The proposed network exploits the residual learning by minimizing the energy-based objective function with powerful image regularization and convex optimization techniques. We demonstrate our proposed approach in quantitative and qualitative experiments that generalize robustly to real input and it is easy to deploy for other down-scaling operators and mobile/embedded devices.

Enkhtogtokh Togootogtokh, Christian Micheloni, Gian Luca Foresti et al.

Modern Unmanned Aerial Vehicles equipped with state of the art artificial intelligence (AI) technologies are opening to a wide plethora of novel and interesting applications. While this field received a strong impact from the recent AI breakthroughs, most of the provided solutions either entirely rely on commercial software or provide a weak integration interface which denies the development of additional techniques. This leads us to propose a novel and efficient framework for the UAV-AI joint technology. Intelligent UAV systems encounter complex challenges to be tackled without human control. One of these complex challenges is to be able to carry out computer vision tasks in real-time use cases. In this paper we focus on this challenge and introduce a multi-layer AI (MLAI) framework to allow easy integration of ad-hoc visual-based AI applications. To show its features and its advantages, we implemented and evaluated different modern visual-based deep learning models for object detection, target tracking and target handover.

Marco Zamprogno, Marco Passon, Niki Martinel et al.

In this paper we consider the problem of video-based person re-identification, which is the task of associating videos of the same person captured by different and non-overlapping cameras. We propose a Siamese framework in which video frames of the person to re-identify and of the candidate one are processed by two identical networks which produce a similarity score. We introduce an attention mechanisms to capture the relevant information both at frame level (spatial information) and at video level (temporal information given by the importance of a specific frame within the sequence). One of the novelties of our approach is given by a joint concurrent processing of both frame and video levels, providing in such a way a very simple architecture. Despite this fact, our approach achieves better performance than the state-of-the-art on the challenging iLIDS-VID dataset.

Matteo Dunnhofer, Niki Martinel, Gian Luca Foresti et al.

In the last decade many different algorithms have been proposed to track a generic object in videos. Their execution on recent large-scale video datasets can produce a great amount of various tracking behaviours. New trends in Reinforcement Learning showed that demonstrations of an expert agent can be efficiently used to speed-up the process of policy learning. Taking inspiration from such works and from the recent applications of Reinforcement Learning to visual tracking, we propose two novel trackers, A3CT, which exploits demonstrations of a state-of-the-art tracker to learn an effective tracking policy, and A3CTD, that takes advantage of the same expert tracker to correct its behaviour during tracking. Through an extensive experimental validation on the GOT-10k, OTB-100, LaSOT, UAV123 and VOT benchmarks, we show that the proposed trackers achieve state-of-the-art performance while running in real-time.

Rao Muhammad Umer, Gian Luca Foresti, Christian Micheloni

Single Image Super-Resolution (SISR) aims to generate a high-resolution (HR) image of a given low-resolution (LR) image. The most of existing convolutional neural network (CNN) based SISR methods usually take an assumption that a LR image is only bicubicly down-sampled version of an HR image. However, the true degradation (i.e. the LR image is a bicubicly downsampled, blurred and noisy version of an HR image) of a LR image goes beyond the widely used bicubic assumption, which makes the SISR problem highly ill-posed nature of inverse problems. To address this issue, we propose a deep SISR network that works for blur kernels of different sizes, and different noise levels in an unified residual CNN-based denoiser network, which significantly improves a practical CNN-based super-resolver for real applications. Extensive experimental results on synthetic LR datasets and real images demonstrate that our proposed method not only can produce better results on more realistic degradation but also computational efficient to practical SISR applications.

Niki Martinel, Gian Luca Foresti, Christian Micheloni

Food diary applications represent a tantalizing market. Such applications, based on image food recognition, opened to new challenges for computer vision and pattern recognition algorithms. Recent works in the field are focusing either on hand-crafted representations or on learning these by exploiting deep neural networks. Despite the success of such a last family of works, these generally exploit off-the shelf deep architectures to classify food dishes. Thus, the architectures are not cast to the specific problem. We believe that better results can be obtained if the deep architecture is defined with respect to an analysis of the food composition. Following such an intuition, this work introduces a new deep scheme that is designed to handle the food structure. Specifically, inspired by the recent success of residual deep network, we exploit such a learning scheme and introduce a slice convolution block to capture the vertical food layers. Outputs of the deep residual blocks are combined with the sliced convolution to produce the classification score for specific food categories. To evaluate our proposed architecture we have conducted experimental results on three benchmark datasets. Results demonstrate that our solution shows better performance with respect to existing approaches (e.g., a top-1 accuracy of 90.27% on the Food-101 challenging dataset).

Niki Martinel, Abir Das, Christian Micheloni et al.

Person re-identification is an open and challenging problem in computer vision. Majority of the efforts have been spent either to design the best feature representation or to learn the optimal matching metric. Most approaches have neglected the problem of adapting the selected features or the learned model over time. To address such a problem, we propose a temporal model adaptation scheme with human in the loop. We first introduce a similarity-dissimilarity learning method which can be trained in an incremental fashion by means of a stochastic alternating directions methods of multipliers optimization procedure. Then, to achieve temporal adaptation with limited human effort, we exploit a graph-based approach to present the user only the most informative probe-gallery matches that should be used to update the model. Results on three datasets have shown that our approach performs on par or even better than state-of-the-art approaches while reducing the manual pairwise labeling effort by about 80%.