Eren Erdal Aksoy

Felix Rosberg, Eren Erdal Aksoy, Fernando Alonso-Fernandez et al.

In this work, we present a new single-stage method for subject agnostic face swapping and identity transfer, named FaceDancer. We have two major contributions: Adaptive Feature Fusion Attention (AFFA) and Interpreted Feature Similarity Regularization (IFSR). The AFFA module is embedded in the decoder and adaptively learns to fuse attribute features and features conditioned on identity information without requiring any additional facial segmentation process. In IFSR, we leverage the intermediate features in an identity encoder to preserve important attributes such as head pose, facial expression, lighting, and occlusion in the target face, while still transferring the identity of the source face with high fidelity. We conduct extensive quantitative and qualitative experiments on various datasets and show that the proposed FaceDancer outperforms other state-of-the-art networks in terms of identityn transfer, while having significantly better pose preservation than most of the previous methods.

Nour Alhuda Albashir, Lars Pernickel, Danial Hamoud et al.

Autonomous vehicles face major perception and navigation challenges in adverse weather such as rain, fog, and snow, which degrade the performance of LiDAR, RADAR, and RGB camera sensors. While each sensor type offers unique strengths, such as RADAR robustness in poor visibility and LiDAR precision in clear conditions, they also suffer distinct limitations when exposed to environmental obstructions. This study proposes LRC-WeatherNet, a novel multi-sensor fusion framework that integrates LiDAR, RADAR, and camera data for real-time classification of weather conditions. By employing both early fusion using a unified Bird's Eye View representation and mid-level gated fusion of modality-specific feature maps, our approach adapts to the varying reliability of each sensor under changing weather. Evaluated on the extensive MSU-4S dataset covering nine weather types, LRC-WeatherNet achieves superior classification performance and computational efficiency, significantly outperforming unimodal baselines in adverse conditions. This work is the first to combine all three modalities for robust, real-time weather classification in autonomous driving. We release our trained models and source code in https://github.com/nouralhudaalbashir/LRC-WeatherNet.

Felix Rosberg, Eren Erdal Aksoy, Cristofer Englund et al.

In this paper, we present a new approach for facial anonymization in images and videos, abbreviated as FIVA. Our proposed method is able to maintain the same face anonymization consistently over frames with our suggested identity-tracking and guarantees a strong difference from the original face. FIVA allows for 0 true positives for a false acceptance rate of 0.001. Our work considers the important security issue of reconstruction attacks and investigates adversarial noise, uniform noise, and parameter noise to disrupt reconstruction attacks. In this regard, we apply different defense and protection methods against these privacy threats to demonstrate the scalability of FIVA. On top of this, we also show that reconstruction attack models can be used for detection of deep fakes. Last but not least, we provide experimental results showing how FIVA can even enable face swapping, which is purely trained on a single target image.

Tiago Cortinhal, Idriss Gouigah, Eren Erdal Aksoy

Although LiDAR sensors are crucial for autonomous systems due to providing precise depth information, they struggle with capturing fine object details, especially at a distance, due to sparse and non-uniform data. Recent advances introduced pseudo-LiDAR, i.e., synthetic dense point clouds, using additional modalities such as cameras to enhance 3D object detection. We present a novel LiDAR-only framework that augments raw scans with denser pseudo point clouds by solely relying on LiDAR sensors and scene semantics, omitting the need for cameras. Our framework first utilizes a segmentation model to extract scene semantics from raw point clouds, and then employs a multi-modal domain translator to generate synthetic image segments and depth cues without real cameras. This yields a dense pseudo point cloud enriched with semantic information. We also introduce a new semantically guided projection method, which enhances detection performance by retaining only relevant pseudo points. We applied our framework to different advanced 3D object detection methods and reported up to 2.9% performance upgrade. We also obtained comparable results on the KITTI 3D object detection dataset, in contrast to other state-of-the-art LiDAR-only detectors.

Tiago Cortinhal, Eren Erdal Aksoy

This work presents a new depth- and semantics-aware conditional generative model, named TITAN-Next, for cross-domain image-to-image translation in a multi-modal setup between LiDAR and camera sensors. The proposed model leverages scene semantics as a mid-level representation and is able to translate raw LiDAR point clouds to RGB-D camera images by solely relying on semantic scene segments. We claim that this is the first framework of its kind and it has practical applications in autonomous vehicles such as providing a fail-safe mechanism and augmenting available data in the target image domain. The proposed model is evaluated on the large-scale and challenging Semantic-KITTI dataset, and experimental findings show that it considerably outperforms the original TITAN-Net and other strong baselines by 23.7$\%$ margin in terms of IoU.

Gamze Akyol, Sanem Sariel, Eren Erdal Aksoy

Despite decades of research, understanding human manipulation activities is, and has always been, one of the most attractive and challenging research topics in computer vision and robotics. Recognition and prediction of observed human manipulation actions have their roots in the applications related to, for instance, human-robot interaction and robot learning from demonstration. The current research trend heavily relies on advanced convolutional neural networks to process the structured Euclidean data, such as RGB camera images. These networks, however, come with immense computational complexity to be able to process high dimensional raw data. Different from the related works, we here introduce a deep graph autoencoder to jointly learn recognition and prediction of manipulation tasks from symbolic scene graphs, instead of relying on the structured Euclidean data. Our network has a variational autoencoder structure with two branches: one for identifying the input graph type and one for predicting the future graphs. The input of the proposed network is a set of semantic graphs which store the spatial relations between subjects and objects in the scene. The network output is a label set representing the detected and predicted class types. We benchmark our new model against different state-of-the-art methods on two different datasets, MANIAC and MSRC-9, and show that our proposed model can achieve better performance. We also release our source code https://github.com/gamzeakyol/GNet.

Arda Inceoglu, Eren Erdal Aksoy, Abdullah Cihan Ak et al.

Safe manipulation in unstructured environments for service robots is a challenging problem. A failure detection system is needed to monitor and detect unintended outcomes. We propose FINO-Net, a novel multimodal sensor fusion based deep neural network to detect and identify manipulation failures. We also introduce a multimodal dataset, containing 229 real-world manipulation data recorded with a Baxter robot. Our network combines RGB, depth and audio readings to effectively detect and classify failures. Results indicate that fusing RGB with depth and audio modalities significantly improves the performance. FINO-Net achieves 98.60% detection and 87.31% classification accuracy on our novel dataset. Code and data are publicly available at https://github.com/ardai/fino-net.

Tiago Cortinhal, George Tzelepis, Eren Erdal Aksoy

In this paper, we introduce SalsaNext for the uncertainty-aware semantic segmentation of a full 3D LiDAR point cloud in real-time. SalsaNext is the next version of SalsaNet [1] which has an encoder-decoder architecture where the encoder unit has a set of ResNet blocks and the decoder part combines upsampled features from the residual blocks. In contrast to SalsaNet, we introduce a new context module, replace the ResNet encoder blocks with a new residual dilated convolution stack with gradually increasing receptive fields and add the pixel-shuffle layer in the decoder. Additionally, we switch from stride convolution to average pooling and also apply central dropout treatment. To directly optimize the Jaccard index, we further combine the weighted cross-entropy loss with Lovasz-Softmax loss [2]. We finally inject a Bayesian treatment to compute the epistemic and aleatoric uncertainties for each point in the cloud. We provide a thorough quantitative evaluation on the Semantic-KITTI dataset [3], which demonstrates that the proposed SalsaNext outperforms other state-of-the-art semantic segmentation networks and ranks first on the Semantic-KITTI leaderboard. We also release our source code https://github.com/TiagoCortinhal/SalsaNext.

Abu Mohammed Raisuddin, Jesper Holmblad, Hamed Haghighi et al.

Sensor degradation poses a significant challenge in autonomous driving. During heavy rainfall, the interference from raindrops can adversely affect the quality of LiDAR point clouds, resulting in, for instance, inaccurate point measurements. This, in turn, can potentially lead to safety concerns if autonomous driving systems are not weather-aware, i.e., if they are unable to discern such changes. In this study, we release a new, large-scale, multi-modal emulated rain dataset, REHEARSE-3D, to promote research advancements in 3D point cloud de-raining. Distinct from the most relevant competitors, our dataset is unique in several respects. First, it is the largest point-wise annotated dataset, and second, it is the only one with high-resolution LiDAR data (LiDAR-256) enriched with 4D Radar point clouds logged in both daytime and nighttime conditions in a controlled weather environment. Furthermore, REHEARSE-3D involves rain-characteristic information, which is of significant value not only for sensor noise modeling but also for analyzing the impact of weather at a point level. Leveraging REHEARSE-3D, we benchmark raindrop detection and removal in fused LiDAR and 4D Radar point clouds. Our comprehensive study further evaluates the performance of various statistical and deep-learning models. Upon publication, the dataset and benchmark models will be made publicly available at: https://sporsho.github.io/REHEARSE3D.

Enes Erdogan, Eren Erdal Aksoy, Sanem Sariel

Recognition of human manipulation actions in real-time is essential for safe and effective human-robot interaction and collaboration. The challenge lies in developing a model that is both lightweight enough for real-time execution and capable of generalization. While some existing methods in the literature can run in real-time, they struggle with temporal scalability, i.e., they fail to adapt to long-duration manipulations effectively. To address this, leveraging the generalizable scene graph representations, we propose a new Factorized Graph Sequence Encoder network that not only runs in real-time but also scales effectively in the temporal dimension, thanks to its factorized encoder architecture. Additionally, we introduce Hand Pooling operation, a simple pooling operation for more focused extraction of the graph-level embeddings. Our model outperforms the previous state-of-the-art real-time approach, achieving a 14.3\% and 5.6\% improvement in F1-macro score on the KIT Bimanual Action (Bimacs) Dataset and Collaborative Action (CoAx) Dataset, respectively. Moreover, we conduct an extensive ablation study to validate our network design choices. Finally, we compare our model with its architecturally similar RGB-based model on the Bimacs dataset and show the limitations of this model in contrast to ours on such an object-centric manipulation dataset.

Georgies Tzelepis, Eren Erdal Aksoy, Júlia Borràs et al.

Understanding of deformable object manipulations such as textiles is a challenge due to the complexity and high dimensionality of the problem. Particularly, the lack of a generic representation of semantic states (e.g., \textit{crumpled}, \textit{diagonally folded}) during a continuous manipulation process introduces an obstacle to identify the manipulation type. In this paper, we aim to solve the problem of semantic state estimation in cloth manipulation tasks. For this purpose, we introduce a new large-scale fully-annotated RGB image dataset showing various human demonstrations of different complicated cloth manipulations. We provide a set of baseline deep networks and benchmark them on the problem of semantic state estimation using our proposed dataset. Furthermore, we investigate the scalability of our semantic state estimation framework in robot monitoring tasks of long and complex cloth manipulations.

Nesma M. Rezk, Tomas Nordström, Dimitrios Stathis et al.

The compression of deep learning models is of fundamental importance in deploying such models to edge devices. The selection of compression parameters can be automated to meet changes in the hardware platform and application using optimization algorithms. This article introduces a Multi-Objective Hardware-Aware Quantization (MOHAQ) method, which considers hardware efficiency and inference error as objectives for mixed-precision quantization. The proposed method feasibly evaluates candidate solutions in a large search space by relying on two steps. First, post-training quantization is applied for fast solution evaluation (inference-only search). Second, we propose the "beacon-based search" to retrain selected solutions only and use them as beacons to know the effect of retraining on other solutions. We use a speech recognition model based on Simple Recurrent Unit (SRU) using the TIMIT dataset and apply our method to run on SiLago and Bitfusion platforms. We provide experimental evaluations showing that SRU can be compressed up to 8x by post-training quantization without any significant error increase. On SiLago, we found solutions that achieve 97\% and 86\% of the maximum possible speedup and energy saving, with a minor increase in error. On Bitfusion, beacon-based search reduced the error gain of inference-only search by up to 4.9 percentage points.

Cristofer Englund, Eren Erdal Aksoy, Fernando Alonso-Fernandez et al.

Smart Cities and Communities (SCC) constitute a new paradigm in urban development. SCC ideates on a data-centered society aiming at improving efficiency by automating and optimizing activities and utilities. Information and communication technology along with internet of things enables data collection and with the help of artificial intelligence (AI) situation awareness can be obtained to feed the SCC actors with enriched knowledge. This paper describes AI perspectives in SCC and gives an overview of AI-based technologies used in traffic to enable road vehicle automation and smart traffic control. Perception, Smart Traffic Control and Driver Modelling are described along with open research challenges and standardization to help introduce advanced driver assistance systems and automated vehicle functionality in traffic. To fully realize the potential of SCC, to create a holistic view on a city level, the availability of data from different stakeholders is need. Further, though AI technologies provide accurate predictions and classifications there is an ambiguity regarding the correctness of their outputs. This can make it difficult for the human operator to trust the system. Today there are no methods that can be used to match function requirements with the level of detail in data annotation in order to train an accurate model. Another challenge related to trust is explainability, while the models have difficulties explaining how they come to a certain conclusion it is difficult for humans to trust it.

Georgios Tzelepis, Ahraz Asif, Saimir Baci et al.

Neural networks have been notorious for being computationally expensive. This is mainly because neural networks are often over-parametrized and most likely have redundant nodes or layers as they are getting deeper and wider. Their demand for hardware resources prohibits their extensive use in embedded devices and puts restrictions on tasks like real-time image classification or object detection. In this work, we propose a network-agnostic model compression method infused with a novel dynamical clustering approach to reduce the computational cost and memory footprint of deep neural networks. We evaluated our new compression method on five different state-of-the-art image classification and object detection networks. In classification networks, we pruned about 95% of network parameters. In advanced detection networks such as YOLOv3, our proposed compression method managed to reduce the model parameters up to 59.70% which yielded 110X less memory without sacrificing much in accuracy.

Martin Cooney, Jacob Pihl, Hanna Larsson et al.

Socially assistive robots are increasingly being designed to interact with humans in various therapeutical scenarios. We believe that one useful scenario is providing exercise coaching for Persons with Dementia (PwD), which involves unique challenges related to memory and communication. We present a design for a robot that can seek to help a PWD to conduct exercises by recognizing behaviors and providing feedback, in an online, multimodal, and engaging way. Additionally, in line with a mid-fidelity prototyping approach, we report on some feedback from an exploratory user study using a Baxter robot, which provided some confirmation of the usefulness of the general scenario; furthermore, the results suggested the degree of a robot's feedback could be tailored to provide impressions of attentiveness or fun. Limitations and possibilities for future improvement are outlined, touching on deep learning and haptic feedback, toward informing next designs.

Eren Erdal Aksoy, Saimir Baci, Selcuk Cavdar

In this paper, we introduce a deep encoder-decoder network, named SalsaNet, for efficient semantic segmentation of 3D LiDAR point clouds. SalsaNet segments the road, i.e. drivable free-space, and vehicles in the scene by employing the Bird-Eye-View (BEV) image projection of the point cloud. To overcome the lack of annotated point cloud data, in particular for the road segments, we introduce an auto-labeling process which transfers automatically generated labels from the camera to LiDAR. We also explore the role of imagelike projection of LiDAR data in semantic segmentation by comparing BEV with spherical-front-view projection and show that SalsaNet is projection-agnostic. We perform quantitative and qualitative evaluations on the KITTI dataset, which demonstrate that the proposed SalsaNet outperforms other state-of-the-art semantic segmentation networks in terms of accuracy and computation time. Our code and data are publicly available at https://gitlab.com/aksoyeren/salsanet.git.

Fabio Ferreira, Jonas Rothfuss, Eren Erdal Aksoy et al.

We release two artificial datasets, Simulated Flying Shapes and Simulated Planar Manipulator that allow to test the learning ability of video processing systems. In particular, the dataset is meant as a tool which allows to easily assess the sanity of deep neural network models that aim to encode, reconstruct or predict video frame sequences. The datasets each consist of 90000 videos. The Simulated Flying Shapes dataset comprises scenes showing two objects of equal shape (rectangle, triangle and circle) and size in which one object approaches its counterpart. The Simulated Planar Manipulator shows a 3-DOF planar manipulator that executes a pick-and-place task in which it has to place a size-varying circle on a squared platform. Different from other widely used datasets such as moving MNIST [1], [2], the two presented datasets involve goal-oriented tasks (e.g. the manipulator grasping an object and placing it on a platform), rather than showing random movements. This makes our datasets more suitable for testing prediction capabilities and the learning of sophisticated motions by a machine learning model. This technical document aims at providing an introduction into the usage of both datasets.

Jonas Rothfuss, Fabio Ferreira, Eren Erdal Aksoy et al.

We present a novel deep neural network architecture for representing robot experiences in an episodic-like memory which facilitates encoding, recalling, and predicting action experiences. Our proposed unsupervised deep episodic memory model 1) encodes observed actions in a latent vector space and, based on this latent encoding, 2) infers most similar episodes previously experienced, 3) reconstructs original episodes, and 4) predicts future frames in an end-to-end fashion. Results show that conceptually similar actions are mapped into the same region of the latent vector space. Based on these results, we introduce an action matching and retrieval mechanism, benchmark its performance on two large-scale action datasets, 20BN-something-something and ActivityNet and evaluate its generalization capability in a real-world scenario on a humanoid robot.

Eren Erdal Aksoy, Adil Orhan, Florentin Woergoetter

Understanding continuous human actions is a non-trivial but important problem in computer vision. Although there exists a large corpus of work in the recognition of action sequences, most approaches suffer from problems relating to vast variations in motions, action combinations, and scene contexts. In this paper, we introduce a novel method for semantic segmentation and recognition of long and complex manipulation action tasks, such as "preparing a breakfast" or "making a sandwich". We represent manipulations with our recently introduced "Semantic Event Chain" (SEC) concept, which captures the underlying spatiotemporal structure of an action invariant to motion, velocity, and scene context. Solely based on the spatiotemporal interactions between manipulated objects and hands in the extracted SEC, the framework automatically parses individual manipulation streams performed either sequentially or concurrently. Using event chains, our method further extracts basic primitive elements of each parsed manipulation. Without requiring any prior object knowledge, the proposed framework can also extract object-like scene entities that exhibit the same role in semantically similar manipulations. We conduct extensive experiments on various recent datasets to validate the robustness of the framework.

Yezhou Yang, Yiannis Aloimonos, Cornelia Fermuller et al.

In this paper we present a formal computational framework for modeling manipulation actions. The introduced formalism leads to semantics of manipulation action and has applications to both observing and understanding human manipulation actions as well as executing them with a robotic mechanism (e.g. a humanoid robot). It is based on a Combinatory Categorial Grammar. The goal of the introduced framework is to: (1) represent manipulation actions with both syntax and semantic parts, where the semantic part employs $λ$-calculus; (2) enable a probabilistic semantic parsing schema to learn the $λ$-calculus representation of manipulation action from an annotated action corpus of videos; (3) use (1) and (2) to develop a system that visually observes manipulation actions and understands their meaning while it can reason beyond observations using propositional logic and axiom schemata. The experiments conducted on a public available large manipulation action dataset validate the theoretical framework and our implementation.