References of "Deliège, Adrien"
     in
Bookmark and Share    
Full Text
Peer Reviewed
See detailA Context-Aware Loss Function for Action Spotting in Soccer Videos
Cioppa, Anthony ULiege; Deliège, Adrien ULiege; Giancola, Silvio et al

in IEEE Conference on Computer Vision and Pattern Recognition. Proceedings (2020, June)

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 ... [more ▼]

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. [less ▲]

Detailed reference viewed: 141 (41 ULiège)
Full Text
Peer Reviewed
See detailMultimodal and multiview distillation for real-time player detection on a football field
Cioppa, Anthony ULiege; Deliège, Adrien ULiege; Noor, Ul Huda et al

in IEEE Conference on Computer Vision and Pattern Recognition. Proceedings (2020, June)

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 ... [more ▼]

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. [less ▲]

Detailed reference viewed: 71 (21 ULiège)
Full Text
See detailDes mathématiques derrière l'intelligence artificielle
Deliège, Adrien ULiege

Conference given outside the academic context (2020)

Présentation donnée dans le cadre du printemps de sciences le 11 mars 2020, à l'occasion du Pi Day, afin de vulgariser les mathématiques pour des élèves de secondaire.

Detailed reference viewed: 46 (1 ULiège)
Full Text
Peer Reviewed
See detailGhost Loss to Question the Reliability of Training Data
Deliège, Adrien ULiege; Cioppa, Anthony ULiege; Van Droogenbroeck, Marc ULiege

in IEEE Access (2020), 8

Supervised image classification problems rely on training data assumed to have been correctly annotated; this assumption underpins most works in the field of deep learning. In consequence, during its ... [more ▼]

Supervised image classification problems rely on training data assumed to have been correctly annotated; this assumption underpins most works in the field of deep learning. In consequence, during its training, a network is forced to match the label provided by the annotator and is not given the flexibility to choose an alternative to inconsistencies that it might be able to detect. Therefore, erroneously labeled training images may end up “correctly” classified in classes which they do not actually belong to. This may reduce the performances of the network and thus incite to build more complex networks without even checking the quality of the training data. In this work, we question the reliability of the annotated datasets. For that purpose, we introduce the notion of ghost loss, which can be seen as a regular loss that is zeroed out for some predicted values in a deterministic way and that allows the network to choose an alternative to the given label without being penalized. After a proof of concept experiment, we use the ghost loss principle to detect confusing images and erroneously labeled images in well-known training datasets (MNIST, Fashion-MNIST, SVHN, CIFAR10) and we provide a new tool, called sanity matrix, for summarizing these confusions. [less ▲]

Detailed reference viewed: 43 (12 ULiège)
Full Text
Peer Reviewed
See detailOrdinal Pooling
Deliège, Adrien ULiege; Istasse, Maxime; Kumar, Ashwani et al

in 30th British Machine Vision Conference (2020)

In the framework of convolutional neural networks, downsampling is often performed with an average-pooling, where all the activations are treated equally, or with a max-pooling operation that only retains ... [more ▼]

In the framework of convolutional neural networks, downsampling is often performed with an average-pooling, where all the activations are treated equally, or with a max-pooling operation that only retains an element with maximum activation while discarding the others. Both of these operations are restrictive and have previously been shown to be sub-optimal. To address this issue, a novel pooling scheme, named ordinal pooling, is introduced in this work. Ordinal pooling rearranges all the elements of a pooling region in a sequence and assigns a different weight to each element based upon its order in the sequence. These weights are used to compute the pooling operation as a weighted sum of the rearranged elements of the pooling region. They are learned via a standard gradient-based training, allowing to learn a behavior anywhere in the spectrum of average-pooling to max-pooling in a differentiable manner. Our experiments suggest that it is advantageous for the networks to perform different types of pooling operations within a pooling layer and that a hybrid behavior between average- and max-pooling is often beneficial. More importantly, they also demonstrate that ordinal pooling leads to consistent improvements in the accuracy over average- or max-pooling operations while speeding up the training and alleviating the issue of the choice of the pooling operations and activation functions to be used in the networks. In particular, ordinal pooling mainly helps on lightweight or quantized deep learning architectures, as typically considered e.g. for embedded applications. [less ▲]

Detailed reference viewed: 119 (11 ULiège)
Full Text
Peer Reviewed
See detailAnalysis of the Topographic Roughness of the Moon Using the Wavelet Leaders Method and the Lunar Digital Elevation Model From the Lunar Orbiter Laser Altimeter and SELENE Terrain Camera
Lemelin, Myriam; Daly, Mike; Deliège, Adrien ULiege

in Journal of Geophysical Research. Planets (2020), 125(1),

The Wavelet Leaders Method (WLM) is a wavelet‐based multifractal formalism that allows the identification of scale breaks (thus scaling regimes), the definition of scaling properties (mono versus multi ... [more ▼]

The Wavelet Leaders Method (WLM) is a wavelet‐based multifractal formalism that allows the identification of scale breaks (thus scaling regimes), the definition of scaling properties (mono versus multi fractality of the surface), and the calculation of the Hölder exponent that characterizes each pixel, based on the comparison between a theoretical wavelet and topographic values. Here we use the WLM and the SLDEM2015 digital elevation model to provide a near‐global and a local isotropic characterization of the lunar roughness. The near‐global study of baselines between 330 m and 1,350 km reveals scale breaks at ~1.3, 42.2, and 337.6 km. Scaling properties and Hölder exponent values were calculated for the three corresponding scaling regimes: 330–659 m, 1.3–21.1 km, and 42.2–168.8 km. We find that the dichotomy between the highlands and the maria is present at all scales. Between 330 and 659 m, the Hölder exponent map shows the unique signature of Orientale basin, rilles, and a correlation with the age of mare units. Between 1.3 and 21.1 km, it shows the unique signature of the Orientale basin and a relationship with the density of 5‐ to 20‐km‐diameter craters. Scaling properties and Hölder exponent values were also calculated locally for complex craters, basins, rilles and light plains, for two scaling regimes: 165–659 m and 1.3–21.1 km. Relationships between the Hölder exponent values at 165–659 m, the density of <500‐m‐diameter craters and different geologic units were found and a potential scale break near 165 m was identified. [less ▲]

Detailed reference viewed: 26 (2 ULiège)
Full Text
See detailImage classification using neural networks
Van Droogenbroeck, Marc ULiege; Deliège, Adrien ULiege; Cioppa, Anthony ULiege

Patent (2019)

A computer-implemented method for training a neural network for classifying image data and a related computer program product are disclosed. A labelled input data set comprising a plurality of labelled ... [more ▼]

A computer-implemented method for training a neural network for classifying image data and a related computer program product are disclosed. A labelled input data set comprising a plurality of labelled image data samples is provided together with a neural network. The neural network comprises an input layer, at least one intermediate layer, and an output layer having one channel per label class. Each channel provides a mapping of labelled image data samples onto feature vectors. Furthermore, the input layer of a decoder network for reconstructing image data samples at its output is connecting the output layer of the neural network. A classifier predicts class labels as the labels of those channels for which a normed distance of its feature vector relative to a pre-determined reference point is smallest. A loss function for the neural network is suitable for steering, for each channel, the feature vectors onto which image data samples of the associated class are mapped, into a convex target region around the pre-determined reference point. [less ▲]

Detailed reference viewed: 113 (33 ULiège)
Full Text
See detailImage classification using neural networks
Van Droogenbroeck, Marc ULiege; Deliège, Adrien ULiege; Cioppa, Anthony ULiege

Patent (2019)

A computer-implemented method for training a neural network for classifying image data and a related computer program product are disclosed. A labelled input data set comprising a plurality of labelled ... [more ▼]

A computer-implemented method for training a neural network for classifying image data and a related computer program product are disclosed. A labelled input data set comprising a plurality of labelled image data samples is provided together with a neural network. The neural network comprises an input layer, at least one intermediate layer, and an output layer having one channel per label class. Each channel provides a mapping of labelled image data samples onto feature vectors. Furthermore, the input layer of a decoder network for reconstructing image data samples at its output is connecting the output layer of the neural network. A classifier predicts class labels as the labels of those channels for which a normed distance of its feature vector relative to a pre-determined reference point is smallest. A loss function for the neural network is suitable for steering, for each channel, the feature vectors onto which image data samples of the associated class are mapped, into a convex target region around the pre-determined reference point. [less ▲]

Detailed reference viewed: 33 (2 ULiège)
Full Text
Peer Reviewed
See detailARTHuS: Adaptive Real-Time Human Segmentation in Sports through Online Distillation
Cioppa, Anthony ULiege; Deliège, Adrien ULiege; Istasse, Maxime et al

in IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) Proceedings (2019, June)

Semantic segmentation can be regarded as a useful tool for global scene understanding in many areas, including sports, but has inherent difficulties, such as the need for pixel-wise annotated training ... [more ▼]

Semantic segmentation can be regarded as a useful tool for global scene understanding in many areas, including sports, but has inherent difficulties, such as the need for pixel-wise annotated training data and the absence of well-performing real-time universal algorithms. To alleviate these issues, we sacrifice universality by developing a general method, named ARTHuS, that produces adaptive real-time match-specific networks for human segmentation in sports videos, without requiring any manual annotation. This is done by an online knowledge distillation process, in which a fast student network is trained to mimic the output of an existing slow but effective universal teacher network, while being periodically updated to adjust to the latest play conditions. As a result, ARTHuS allows to build highly effective real-time human segmentation networks that evolve through the match and that sometimes outperform their teacher. The usefulness of producing adaptive match-specific networks and their excellent performances are demonstrated quantitatively and qualitatively for soccer and basketball matches. [less ▲]

Detailed reference viewed: 1023 (123 ULiège)
Full Text
See detailDes mathématiques derrière l'intelligence artificielle
Deliège, Adrien ULiege

Conference given outside the academic context (2019)

Detailed reference viewed: 68 (8 ULiège)
Full Text
Peer Reviewed
See detailAn Effective Hit-or-Miss Layer Favoring Feature Interpretation as Learned Prototypes Deformations
Deliège, Adrien ULiege; Cioppa, Anthony ULiege; Van Droogenbroeck, Marc ULiege

in Thirty-Third AAAI Conference on Artificial Intelligence (2019, February)

Neural networks designed for the task of classification have become a commodity in recent years. Many works target the development of more effective networks, which results in a complexification of their ... [more ▼]

Neural networks designed for the task of classification have become a commodity in recent years. Many works target the development of more effective networks, which results in a complexification of their architectures with more layers, multiple sub-networks, or even the combination of multiple classifiers, but this often comes at the expense of producing uninterpretable black boxes. In this paper, we redesign a simple capsule network to enable it to synthesize class-representative samples, called prototypes, by replacing the last layer with a novel Hit-or-Miss layer. This layer contains activated vectors, called capsules, that we train to hit or miss a fixed target capsule by tailoring a specific centripetal loss function. This possibility allows to develop a data augmentation step combining information from the data space and the feature space, resulting in a hybrid data augmentation process. We show that our network, named HitNet, is able to reach better performances than those reproduced with the initial CapsNet on several datasets, while allowing to visualize the nature of the features extracted as deformations of the prototypes, which provides a direct insight into the feature representation learned by the network. [less ▲]

Detailed reference viewed: 337 (88 ULiège)
Full Text
See detailHitNet: a neural network with capsules embedded in a Hit-or-Miss layer, extended with hybrid data augmentation and ghost capsules
Deliège, Adrien ULiege; Cioppa, Anthony ULiege; Van Droogenbroeck, Marc ULiege

in arXiv (2018), 1806.06519

Abstract Neural networks designed for the task of classification have become a commodity in recent years. Many works target the development of better networks, which results in a complexification of their ... [more ▼]

Abstract Neural networks designed for the task of classification have become a commodity in recent years. Many works target the development of better networks, which results in a complexification of their architectures with more layers, multiple sub-networks, or even the combination of multiple classifiers. In this paper, we show how to redesign a simple network to reach excellent performances, which are better than the results reproduced with CapsNet on several datasets, by replacing a layer with a Hit-or-Miss layer. This layer contains activated vectors, called capsules, that we train to hit or miss a central capsule by tailoring a specific centripetal loss function. We also show how our network, named HitNet, is capable of synthesizing a representative sample of the images of a given class by including a reconstruction network. This possibility allows to develop a data augmentation step combining information from the data space and the feature space, resulting in a hybrid data augmentation process. In addition, we introduce the possibility for HitNet, to adopt an alternative to the true target when needed by using the new concept of ghost capsules, which is used here to detect potentially mislabeled images in the training data. [less ▲]

Detailed reference viewed: 109 (40 ULiège)
Full Text
Peer Reviewed
See detailEvidence for solar influence in a Holocene speleothem record (Père Noël cave, SE Belgium)
Allan, Mohammed ULiege; Deliège, Adrien ULiege; Verheyden, sophie et al

in Quaternary Science Reviews (2018)

We present a decadal-centennial scale Holocene climate record based on trace elements contents from a 65cm stalagmite from Belgian Père Noël cave. “Père Noël” (PN) stalagmite covers the last 12.7 ka ... [more ▼]

We present a decadal-centennial scale Holocene climate record based on trace elements contents from a 65cm stalagmite from Belgian Père Noël cave. “Père Noël” (PN) stalagmite covers the last 12.7 ka according to U/ Th dating. High spatial resolution measurements of trace elements (Sr, Ba and Mg) were done by Laser-Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS). Trace elements profiles were interpreted as environmental and climate changes in the Han-sur-Lesse region. Power spectrum estimators and continuous wavelet transform were applied on trace elements time series to detect any statistically significant periodicities in the PN stalagmite. Spectral analyses reveal decadal to millennial periodicities (i.e., 68–75, 133–136, 198–209, 291–358, 404–602, 912–1029 and 2365–2670 yr) in the speleothem record. Results were compared to reconstructed sunspot number data to determine whether solar signal is presents in PN speleothem. The occurrence of significant solar periodicities (i.e., cycles of Gleissberg, de Vries, unnamed 500 years, Eddy and Hallstatt) supports for an impact of solar forcing on PN speleothem trace element contents. Moreover, several intervals of significant rapid winter change were detected during the Holocene at 10.3, 9.3–9.5, around 8.2, 6.4–6.2, 4.7–4.5, and around 2.7 ka BP. Those intervals are similar to the cold winter events evidenced in different natural paleoclimate archives, suggesting common climate forcing mechanisms related to changes in solar irradiance. [less ▲]

Detailed reference viewed: 134 (45 ULiège)
Full Text
Peer Reviewed
See detailA bottom-up approach based on semantics for the interpretation of the main camera stream in soccer games
Cioppa, Anthony ULiege; Deliège, Adrien ULiege; Van Droogenbroeck, Marc ULiege

in IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) (2018, June)

Automatic interpretation of sports games is a major challenge, especially when these sports feature complex players organizations and game phases. This paper describes a bottom-up approach based on the ... [more ▼]

Automatic interpretation of sports games is a major challenge, especially when these sports feature complex players organizations and game phases. This paper describes a bottom-up approach based on the extraction of semantic features from the video stream of the main camera in the particular case of soccer using scene-specific techniques. In our approach, all the features, ranging from the pixel level to the game event level, have a semantic meaning. First, we design our own scene-specific deep learning semantic segmentation network and hue histogram analysis to extract pixel-level semantics for the field, players, and lines. These pixel-level semantics are then processed to compute interpretative semantic features which represent characteristics of the game in the video stream that are exploited to interpret soccer. For example, they correspond to how players are distributed in the image or the part of the field that is filmed. Finally, we show how these interpretative semantic features can be used to set up and train a semantic-based decision tree classifier for major game events with a restricted amount of training data. The main advantages of our semantic approach are that it only requires the video feed of the main camera to extract the semantic features, with no need for camera calibration, field homography, player tracking, or ball position estimation. While the automatic interpretation of sports games remains challenging, our approach allows us to achieve promising results for the semantic feature extraction and for the classification between major soccer game events such as attack, goal or goal opportunity, defense, and middle game. [less ▲]

Detailed reference viewed: 362 (71 ULiège)
Full Text
Peer Reviewed
See detailAnalysis and indication on long-term forecasting of the Oceanic Niño index with wavelet-induced components
Deliège, Adrien ULiege; Nicolay, Samuel ULiege

Poster (2018, April)

The present work provides an analysis and a long-term forecasting scheme of the Oceanic Niño Index (ONI) using the continuous wavelet transform. First, it appears that oscillatory components with main ... [more ▼]

The present work provides an analysis and a long-term forecasting scheme of the Oceanic Niño Index (ONI) using the continuous wavelet transform. First, it appears that oscillatory components with main periods of about 17, 31, 43, 61 and 140 months govern most of the variability of the signal, which is consistent with previous works. Then, this information enables us to derive a simple algorithm to model and forecast ONI. The model is based on the observation that the modes extracted from the signal are generally phased with positive or negative anomalies of ONI El Niño and La Niña events). Such a feature is exploited to generate locally stationary curves that mimic this behavior and which can be easily extrapolated to form a basic forecast. The wavelet transform is then used again to smooth out the process and finalize the predictions. The skills of the technique described in this paper are assessed through retroactive forecasts of past El Niño and La Niña events and via classic indicators computed as functions of the lead time. The main asset of the proposed model resides in its long-lead prediction skills. Consequently, this approach should prove helpful as a complement to other models for estimating the long-term trends of ONI. [less ▲]

Detailed reference viewed: 32 (5 ULiège)
Full Text
See detailDes mathématiques derrière l'intelligence artificielle
Deliège, Adrien ULiege

Conference given outside the academic context (2018)

Detailed reference viewed: 69 (12 ULiège)
Full Text
Peer Reviewed
See detailExtracting oscillating components from nonstationary time series: A wavelet-induced method
Deliège, Adrien ULiege; Nicolay, Samuel ULiege

in Physical Review. E, Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics (2017), 96

This paper consists in the description and application of a method called wavelet-induced mode extraction (WIME) in the context of time-frequency analysis. WIME aims to extract the oscillating components ... [more ▼]

This paper consists in the description and application of a method called wavelet-induced mode extraction (WIME) in the context of time-frequency analysis. WIME aims to extract the oscillating components that build amplitude modulated-frequency modulated signals. The essence of this technique relies on the successive extractions of the dominant ridges of wavelet-based time-frequency representations of the signal under consideration. Our tests on simulated examples indicate strong decomposition and reconstruction skills, trouble-free handling of crossing trajectories in the time-frequency plane, sharp performances in frequency detection in the case of mode-mixing problems, and a natural tolerance to noise. These results are compared with those obtained with empirical mode decomposition. We also show that WIME still gives meaningful results with real-life data, namely, the Oceanic Niño Index. [less ▲]

Detailed reference viewed: 156 (51 ULiège)
Full Text
Peer Reviewed
See detailDoes Belgian Holocene speleothem record solar forcing and cold events?
Allan, Mohammed ULiege; Deliège, Adrien ULiege; Nicolay, Samuel ULiege et al

in Climate of the Past Discussions (2017)

We present a decadal-centennial scale Holocene climate record based on trace elements contents from a 65 cm stalagmite (Père Noël) from Belgian Père Noël cave. Père Noël (PN) stalagmite covers the last 12 ... [more ▼]

We present a decadal-centennial scale Holocene climate record based on trace elements contents from a 65 cm stalagmite (Père Noël) from Belgian Père Noël cave. Père Noël (PN) stalagmite covers the last 12.7 ka according to U/Th dating. High spatial resolution measurements of trace elements (Sr, Ba, Mg and Al) were done by Laser-Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS). Trace elements profiles were interpreted as environmental and climate changes in the Han-sur-Lesse region. Power spectrum estimators and continuous wavelet transform were applied on trace elements time series to detect any statistically significant periodicities in the PN stalagmite. Spectral analyses reveal decadal to millennial periodicities (i.e., 68–75, 133–136, 198–209, 291–358, 404–602, 912–1029 and 2365–2670 yr) in the speleothem record. Results were compared to reconstructed sunspot number data to determine whether solar signal is presents in PN speleothem. The occurrence of significant solar periodicities (i.e., cycles of Gleissberg, de Vries, unnamed 500 years, Eddy and Hallstat) supports for an impact of solar forcing on PN speleothem trace elements contents. Moreover, several intervals of significant rapid climate change were detected during the Holocene at 10.3, 9.3–9.5, around 8.2, 6.4–6.2, 4.7–4.5, and around 2.7 ka BP. Those intervals are similar to the cold events evidenced in different natural paleoclimate archivers, suggesting common climate forcing mechanisms related to changes in solar irradiance. [less ▲]

Detailed reference viewed: 128 (15 ULiège)