arXiv subjects

Thomas B. Moeslund

Publications and source records attributed to Thomas B. Moeslund.

69 records · Page 4Linked to original sources

arXiv2021

Dyadformer: A Multi-modal Transformer for Long-Range Modeling of Dyadic Interactions

Personality computing has become an emerging topic in computer vision, due to the wide range of applications it can be used for. However, most works on the topic have focused on analyzing the individual, even when applied to interaction scenarios, and for short periods of time. To address these limitations, we present the Dyadformer, a novel multi-modal multi-subject Transformer architecture to model individual and interpersonal features in dyadic interactions using variable time windows, thus allowing the capture of long-term interdependencies. Our proposed cross-subject layer allows the network to explicitly model interactions among subjects through attentional operations. This proof-of-concept approach shows how multi-modality and joint modeling of both interactants for longer periods of time helps to predict individual attributes. With Dyadformer, we improve state-of-the-art self-reported personality inference results on individual subjects on the UDIVA v0.5 dataset.

cs.CV↗

arXiv2021

Is it Raining Outside? Detection of Rainfall using General-Purpose Surveillance Cameras

In integrated surveillance systems based on visual cameras, the mitigation of adverse weather conditions is an active research topic. Within this field, rain removal algorithms have been developed that artificially remove rain streaks from images or video. In order to deploy such rain removal algorithms in a surveillance setting, one must detect if rain is present in the scene. In this paper, we design a system for the detection of rainfall by the use of surveillance cameras. We reimplement the former state-of-the-art method for rain detection and compare it against a modern CNN-based method by utilizing 3D convolutions. The two methods are evaluated on our new AAU Visual Rain Dataset (VIRADA) that consists of 215 hours of general-purpose surveillance video from two traffic crossings. The results show that the proposed 3D CNN outperforms the previous state-of-the-art method by a large margin on all metrics, for both of the traffic crossings. Finally, it is shown that the choice of region-of-interest has a large influence on performance when trying to generalize the investigated methods. The AAU VIRADA dataset and our implementation of the two rain detection algorithms are publicly available at https://bitbucket.org/aauvap/aau-virada.

cs.CV↗

arXiv2021

SoccerNet-v2: A Dataset and Benchmarks for Holistic Understanding of Broadcast Soccer Videos

Understanding broadcast videos is a challenging task in computer vision, as it requires generic reasoning capabilities to appreciate the content offered by the video editing. In this work, we propose SoccerNet-v2, a novel large-scale corpus of manual annotations for the SoccerNet video dataset, along with open challenges to encourage more research in soccer understanding and broadcast production. Specifically, we release around 300k annotations within SoccerNet's 500 untrimmed broadcast soccer videos. We extend current tasks in the realm of soccer to include action spotting, camera shot segmentation with boundary detection, and we define a novel replay grounding task. For each task, we provide and discuss benchmark results, reproducible with our open-source adapted implementations of the most relevant works in the field. SoccerNet-v2 is presented to the broader research community to help push computer vision closer to automatic solutions for more general video understanding and production purposes.

cs.CV↗

arXiv2021

Sewer-ML: A Multi-Label Sewer Defect Classification Dataset and Benchmark

Perhaps surprisingly sewerage infrastructure is one of the most costly infrastructures in modern society. Sewer pipes are manually inspected to determine whether the pipes are defective. However, this process is limited by the number of qualified inspectors and the time it takes to inspect a pipe. Automatization of this process is therefore of high interest. So far, the success of computer vision approaches for sewer defect classification has been limited when compared to the success in other fields mainly due to the lack of public datasets. To this end, in this work we present a large novel and publicly available multi-label classification dataset for image-based sewer defect classification called Sewer-ML. The Sewer-ML dataset consists of 1.3 million images annotated by professional sewer inspectors from three different utility companies across nine years. Together with the dataset, we also present a benchmark algorithm and a novel metric for assessing performance. The benchmark algorithm is a result of evaluating 12 state-of-the-art algorithms, six from the sewer defect classification domain and six from the multi-label classification domain, and combining the best performing algorithms. The novel metric is a class-importance weighted F2 score, $\text{F}2_{\text{CIW}}$, reflecting the economic impact of each class, used together with the normal pipe F1 score, $\text{F}1_{\text{Normal}}$. The benchmark algorithm achieves an $\text{F}2_{\text{CIW}}$ score of 55.11% and $\text{F}1_{\text{Normal}}$ score of 90.94%, leaving ample room for improvement on the Sewer-ML dataset. The code, models, and dataset are available at the project page https://vap.aau.dk/sewer-ml/

cs.CV↗

arXiv2021

Real-World Super-Resolution of Face-Images from Surveillance Cameras

Most existing face image Super-Resolution (SR) methods assume that the Low-Resolution (LR) images were artificially downsampled from High-Resolution (HR) images with bicubic interpolation. This operation changes the natural image characteristics and reduces noise. Hence, SR methods trained on such data most often fail to produce good results when applied to real LR images. To solve this problem, we propose a novel framework for generation of realistic LR/HR training pairs. Our framework estimates realistic blur kernels, noise distributions, and JPEG compression artifacts to generate LR images with similar image characteristics as the ones in the source domain. This allows us to train a SR model using high quality face images as Ground-Truth (GT). For better perceptual quality we use a Generative Adversarial Network (GAN) based SR model where we have exchanged the commonly used VGG-loss [24] with LPIPS-loss [52]. Experimental results on both real and artificially corrupted face images show that our method results in more detailed reconstructions with less noise compared to existing State-of-the-Art (SoTA) methods. In addition, we show that the traditional non-reference Image Quality Assessment (IQA) methods fail to capture this improvement and demonstrate that the more recent NIMA metric [16] correlates better with human perception via Mean Opinion Rank (MOR).

cs.CV↗

arXiv2020

3D-ZeF: A 3D Zebrafish Tracking Benchmark Dataset

In this work we present a novel publicly available stereo based 3D RGB dataset for multi-object zebrafish tracking, called 3D-ZeF. Zebrafish is an increasingly popular model organism used for studying neurological disorders, drug addiction, and more. Behavioral analysis is often a critical part of such research. However, visual similarity, occlusion, and erratic movement of the zebrafish makes robust 3D tracking a challenging and unsolved problem. The proposed dataset consists of eight sequences with a duration between 15-120 seconds and 1-10 free moving zebrafish. The videos have been annotated with a total of 86,400 points and bounding boxes. Furthermore, we present a complexity score and a novel open-source modular baseline system for 3D tracking of zebrafish. The performance of the system is measured with respect to two detectors: a naive approach and a Faster R-CNN based fish head detector. The system reaches a MOTA of up to 77.6%. Links to the code and dataset is available at the project page https://vap.aau.dk/3d-zef

cs.CV↗

arXiv2020

SIDU: Similarity Difference and Uniqueness Method for Explainable AI

A new brand of technical artificial intelligence ( Explainable AI ) research has focused on trying to open up the 'black box' and provide some explainability. This paper presents a novel visual explanation method for deep learning networks in the form of a saliency map that can effectively localize entire object regions. In contrast to the current state-of-the art methods, the proposed method shows quite promising visual explanations that can gain greater trust of human expert. Both quantitative and qualitative evaluations are carried out on both general and clinical data sets to confirm the effectiveness of the proposed method.

cs.CV↗

arXiv2020

Multimodal and multiview distillation for real-time player detection on a football field

Monitoring the occupancy of public sports facilities is essential to assess their use and to motivate their construction in new places. In the case of a football field, the area to cover is large, thus several regular cameras should be used, which makes the setup expensive and complex. As an alternative, we developed a system that detects players from a unique cheap and wide-angle fisheye camera assisted by a single narrow-angle thermal camera. In this work, we train a network in a knowledge distillation approach in which the student and the teacher have different modalities and a different view of the same scene. In particular, we design a custom data augmentation combined with a motion detection algorithm to handle the training in the region of the fisheye camera not covered by the thermal one. We show that our solution is effective in detecting players on the whole field filmed by the fisheye camera. We evaluate it quantitatively and qualitatively in the case of an online distillation, where the student detects players in real time while being continuously adapted to the latest video conditions.

cs.CV↗

arXiv2020

Evaluation of Model Selection for Kernel Fragment Recognition in Corn Silage

Model selection when designing deep learning systems for specific use-cases can be a challenging task as many options exist and it can be difficult to know the trade-off between them. Therefore, we investigate a number of state of the art CNN models for the task of measuring kernel fragmentation in harvested corn silage. The models are evaluated across a number of feature extractors and image sizes in order to determine optimal model design choices based upon the trade-off between model complexity, accuracy and speed. We show that accuracy improvements can be made with more complex meta-architectures and speed can be optimised by decreasing the image size with only slight losses in accuracy. Additionally, we show improvements in Average Precision at an Intersection over Union of 0.5 of up to 20 percentage points while also decreasing inference time in comparison to previously published work. This result for better model selection enables opportunities for creating systems that can aid farmers in improving their silage quality while harvesting.

cs.CV↗

arXiv2020

A Context-Aware Loss Function for Action Spotting in Soccer Videos

In video understanding, action spotting consists in temporally localizing human-induced events annotated with single timestamps. In this paper, we propose a novel loss function that specifically considers the temporal context naturally present around each action, rather than focusing on the single annotated frame to spot. We benchmark our loss on a large dataset of soccer videos, SoccerNet, and achieve an improvement of 12.8% over the baseline. We show the generalization capability of our loss for generic activity proposals and detection on ActivityNet, by spotting the beginning and the end of each activity. Furthermore, we provide an extended ablation study and display challenging cases for action spotting in soccer videos. Finally, we qualitatively illustrate how our loss induces a precise temporal understanding of actions and show how such semantic knowledge can be used for automatic highlights generation.

cs.CV↗

arXiv2019

A Double-Deep Spatio-Angular Learning Framework for Light Field based Face Recognition

Face recognition has attracted increasing attention due to its wide range of applications, but it is still challenging when facing large variations in the biometric data characteristics. Lenslet light field cameras have recently come into prominence to capture rich spatio-angular information, thus offering new possibilities for advanced biometric recognition systems. This paper proposes a double-deep spatio-angular learning framework for light field based face recognition, which is able to learn both texture and angular dynamics in sequence using convolutional representations; this is a novel recognition framework that has never been proposed before for either face recognition or any other visual recognition task. The proposed double-deep learning framework includes a long short-term memory (LSTM) recurrent network whose inputs are VGG-Face descriptions that are computed using a VGG-Very-Deep-16 convolutional neural network (CNN). The VGG-16 network uses different face viewpoints rendered from a full light field image, which are organised as a pseudo-video sequence. A comprehensive set of experiments has been conducted with the IST-EURECOM light field face database, for varied and challenging recognition tasks. Results show that the proposed framework achieves superior face recognition performance when compared to the state-of-the-art.

cs.CV↗

arXiv2018

Rain Removal in Traffic Surveillance: Does it Matter?

Varying weather conditions, including rainfall and snowfall, are generally regarded as a challenge for computer vision algorithms. One proposed solution to the challenges induced by rain and snowfall is to artificially remove the rain from images or video using rain removal algorithms. It is the promise of these algorithms that the rain-removed image frames will improve the performance of subsequent segmentation and tracking algorithms. However, rain removal algorithms are typically evaluated on their ability to remove synthetic rain on a small subset of images. Currently, their behavior is unknown on real-world videos when integrated with a typical computer vision pipeline. In this paper, we review the existing rain removal algorithms and propose a new dataset that consists of 22 traffic surveillance sequences under a broad variety of weather conditions that all include either rain or snowfall. We propose a new evaluation protocol that evaluates the rain removal algorithms on their ability to improve the performance of subsequent segmentation, instance segmentation, and feature tracking algorithms under rain and snow. If successful, the de-rained frames of a rain removal algorithm should improve segmentation performance and increase the number of accurately tracked features. The results show that a recent single-frame-based rain removal algorithm increases the segmentation performance by 19.7% on our proposed dataset, but it eventually decreases the feature tracking performance and showed mixed results with recent instance segmentation methods. However, the best video-based rain removal algorithm improves the feature tracking accuracy by 7.72%.

cs.CV↗

arXiv2018

The AAU Multimodal Annotation Toolboxes: Annotating Objects in Images and Videos

This tech report gives an introduction to two annotation toolboxes that enable the creation of pixel and polygon-based masks as well as bounding boxes around objects of interest. Both toolboxes support the annotation of sequential images in the RGB and thermal modalities. Each annotated object is assigned a classification tag, a unique ID, and one or more optional meta data tags. The toolboxes are written in C++ with the OpenCV and Qt libraries and are operated by using the visual interface and the extensive range of keyboard shortcuts. Pre-built binaries are available for Windows and MacOS and the tools can be built from source under Linux as well. So far, tens of thousands of frames have been annotated using the toolboxes.

cs.CV↗

arXiv2016

Learning Dynamic Classes of Events using Stacked Multilayer Perceptron Networks

People often use a web search engine to find information about events of interest, for example, sport competitions, political elections, festivals and entertainment news. In this paper, we study a problem of detecting event-related queries, which is the first step before selecting a suitable time-aware retrieval model. In general, event-related information needs can be observed in query streams through various temporal patterns of user search behavior, e.g., spiky peaks for popular events, and periodicities for repetitive events. However, it is also common that users search for non-popular events, which may not exhibit temporal variations in query streams, e.g., past events recently occurred, historical events triggered by anniversaries or similar events, and future events anticipated to happen. To address the challenge of detecting dynamic classes of events, we propose a novel deep learning model to classify a given query into a predetermined set of multiple event types. Our proposed model, a Stacked Multilayer Perceptron (S-MLP) network, consists of multilayer perceptron used as a basic learning unit. We assemble stacked units to further learn complex relationships between neutrons in successive layers. To evaluate our proposed model, we conduct experiments using real-world queries and a set of manually created ground truth. Preliminary results have shown that our proposed deep learning model outperforms the state-of-the-art classification models significantly.

cs.IR↗

arXiv2016

A comprehensive study of sparse codes on abnormality detection

Sparse representation has been applied successfully in abnormal event detection, in which the baseline is to learn a dictionary accompanied by sparse codes. While much emphasis is put on discriminative dictionary construction, there are no comparative studies of sparse codes regarding abnormality detection. We comprehensively study two types of sparse codes solutions - greedy algorithms and convex L1-norm solutions - and their impact on abnormality detection performance. We also propose our framework of combining sparse codes with different detection methods. Our comparative experiments are carried out from various angles to better understand the applicability of sparse codes, including computation time, reconstruction error, sparsity, detection accuracy, and their performance combining various detection methods. Experiments show that combining OMP codes with maximum coordinate detection could achieve state-of-the-art performance on the UCSD dataset [14].

cs.CV↗