Michal Balazia

Philipp Müller, Michal Balazia, Tobias Baur et al.

Automatic analysis of human behaviour is a fundamental prerequisite for the creation of machines that can effectively interact with- and support humans in social interactions. In MultiMediate'23, we address two key human social behaviour analysis tasks for the first time in a controlled challenge: engagement estimation and bodily behaviour recognition in social interactions. This paper describes the MultiMediate'23 challenge and presents novel sets of annotations for both tasks. For engagement estimation we collected novel annotations on the NOvice eXpert Interaction (NOXI) database. For bodily behaviour recognition, we annotated test recordings of the MPIIGroupInteraction corpus with the BBSI annotation scheme. In addition, we present baseline results for both challenge tasks.

Michal Balazia, Philipp Müller, Ákos Levente Tánczos et al.

Body language is an eye-catching social signal and its automatic analysis can significantly advance artificial intelligence systems to understand and actively participate in social interactions. While computer vision has made impressive progress in low-level tasks like head and body pose estimation, the detection of more subtle behaviors such as gesturing, grooming, or fumbling is not well explored. In this paper we present BBSI, the first set of annotations of complex Bodily Behaviors embedded in continuous Social Interactions in a group setting. Based on previous work in psychology, we manually annotated 26 hours of spontaneous human behavior in the MPIIGroupInteraction dataset with 15 distinct body language classes. We present comprehensive descriptive statistics on the resulting dataset as well as results of annotation quality evaluations. For automatic detection of these behaviors, we adapt the Pyramid Dilated Attention Network (PDAN), a state-of-the-art approach for human action detection. We perform experiments using four variants of spatial-temporal features as input to PDAN: Two-Stream Inflated 3D CNN, Temporal Segment Networks, Temporal Shift Module and Swin Transformer. Results are promising and indicate a great room for improvement in this difficult task. Representing a key piece in the puzzle towards automatic understanding of social behavior, BBSI is fully available to the research community.

Tanay Agrawal, Michal Balazia, Philipp Müller et al.

Human behavior understanding requires looking at minute details in the large context of a scene containing multiple input modalities. It is necessary as it allows the design of more human-like machines. While transformer approaches have shown great improvements, they face multiple challenges such as lack of data or background noise. To tackle these, we introduce the Forced Attention (FAt) Transformer which utilize forced attention with a modified backbone for input encoding and a use of additional inputs. In addition to improving the performance on different tasks and inputs, the modification requires less time and memory resources. We provide a model for a generalised feature extraction for tasks concerning social signals and behavior analysis. Our focus is on understanding behavior in videos where people are interacting with each other or talking into the camera which simulates the first person point of view in social interaction. FAt Transformers are applied to two downstream tasks: personality recognition and body language recognition. We achieve state-of-the-art results for Udiva v0.5, First Impressions v2 and MPII Group Interaction datasets. We further provide an extensive ablation study of the proposed architecture.

Utkarsh Tiwari, Snehashis Majhi, Michal Balazia et al.

Video anomaly detection (VAD) in autonomous driving scenario is an important task, however it involves several challenges due to the ego-centric views and moving camera. Due to this, it remains largely under-explored. While recent developments in weakly-supervised VAD methods have shown remarkable progress in detecting critical real-world anomalies in static camera scenario, the development and validation of such methods are yet to be explored for moving camera VAD. This is mainly due to existing datasets like DoTA not following training pre-conditions of weakly-supervised learning. In this paper, we aim to promote weakly-supervised method development for autonomous driving VAD. We reorganize the DoTA dataset and aim to validate recent powerful weakly-supervised VAD methods on moving camera scenarios. Further, we provide a detailed analysis of what modifications on state-of-the-art methods can significantly improve the detection performance. Towards this, we propose a "feature transformation block" and through experimentation we show that our propositions can empower existing weakly-supervised VAD methods significantly in improving the VAD in autonomous driving. Our codes/dataset/demo will be released at github.com/ut21/WSAD-Driving

Michal Balazia, Katerina Hlavackova-Schindler, Petr Sojka et al.

Which joint interactions in the human gait cycle can be used as biometric characteristics? Most current methods on gait recognition suffer from the lack of interpretability. We propose an interpretable feature representation of gait sequences by the graphical Granger causal inference. Gait sequence of a person in the standardized motion capture format, constituting a set of 3D joint spatial trajectories, is envisaged as a causal system of joints interacting in time. We apply the graphical Granger model (GGM) to obtain the so-called Granger causal graph among joints as a discriminative and visually interpretable representation of a person's gait. We evaluate eleven distance functions in the GGM feature space by established classification and class-separability evaluation metrics. Our experiments indicate that, depending on the metric, the most appropriate distance functions for the GGM are the total norm distance and the Ky-Fan 1-norm distance. Experiments also show that the GGM is able to detect the most discriminative joint interactions and that it outperforms five related interpretable models in correct classification rate and in Davies-Bouldin index. The proposed GGM model can serve as a complementary tool for gait analysis in kinesiology or for gait recognition in video surveillance.

Nishit Poddar, Aglind Reka, Diana-Laura Borza et al.

Micro-actions, fleeting and low-amplitude motions, such as glances, nods, or minor posture shifts, carry rich social meaning but remain difficult for current action recognition models to recognize due to their subtlety, short duration, and high inter-class ambiguity. In this paper, we introduce B-MoE, a Body-part-aware Mixture-of-Experts framework designed to explicitly model the structured nature of human motion. In B-MoE, each expert specializes in a distinct body region (head, body, upper limbs, lower limbs), and is based on the lightweight Macro-Micro Motion Encoder (M3E) that captures long-range contextual structure and fine-grained local motion. A cross-attention routing mechanism learns inter-region relationships and dynamically selects the most informative regions for each micro-action. B-MoE uses a dual-stream encoder that fuses these region-specific semantic cues with global motion features to jointly capture spatially localized cues and temporally subtle variations that characterize micro-actions. Experiments on three challenging benchmarks (MA-52, SocialGesture, and MPII-GroupInteraction) show consistent state-of-theart gains, with improvements in ambiguous, underrepresented, and low amplitude classes.

Aglind Reka, Diana Laura Borza, Dominick Reilly et al.

Temporal Action Detection (TAD), the task of localizing and classifying actions in untrimmed video, remains challenging due to action overlaps and variable action durations. Recent findings suggest that TAD performance is dependent on the structural design of transformers rather than on the self-attention mechanism. Building on this insight, we propose a refined feature extraction process through lightweight, yet effective operations. First, we employ a local branch that employs parallel convolutions with varying window sizes to capture both fine-grained and coarse-grained temporal features. This branch incorporates a gating mechanism to select the most relevant features. Second, we introduce a context branch that uses boundary frames as key-value pairs to analyze their relationship with the central frame through cross-attention. The proposed method captures temporal dependencies and improves contextual understanding. Evaluations of the gating mechanism and context branch on challenging datasets (THUMOS14 and EPIC-KITCHEN 100) show a consistent improvement over the baseline and existing methods.

Valeriya Strizhkova, Hadi Kachmar, Hava Chaptoukaev et al.

Human emotions entail a complex set of behavioral, physiological and cognitive changes. Current state-of-the-art models fuse the behavioral and physiological components using classic machine learning, rather than recent deep learning techniques. We propose to fill this gap, designing the Multimodal for Video and Physio (MVP) architecture, streamlined to fuse video and physiological signals. Differently then others approaches, MVP exploits the benefits of attention to enable the use of long input sequences (1-2 minutes). We have studied video and physiological backbones for inputting long sequences and evaluated our method with respect to the state-of-the-art. Our results show that MVP outperforms former methods for emotion recognition based on facial videos, EDA, and ECG/PPG.

Tim Lachmann, Alexandra Israelsson, Christina Tornberg et al.

Humans often experience not just a single basic emotion at a time, but rather a blend of several emotions with varying salience. Despite the importance of such blended emotions, most video-based emotion recognition approaches are designed to recognize single emotions only. The few approaches that have attempted to recognize blended emotions typically cannot assess the relative salience of the emotions within a blend. This limitation largely stems from the lack of datasets containing a substantial number of blended emotion samples annotated with relative salience. To address this shortcoming, we introduce BLEMORE, a novel dataset for multimodal (video, audio) blended emotion recognition that includes information on the relative salience of each emotion within a blend. BLEMORE comprises over 3,000 clips from 58 actors, performing 6 basic emotions and 10 distinct blends, where each blend has 3 different salience configurations (50/50, 70/30, and 30/70). Using this dataset, we conduct extensive evaluations of state-of-the-art video classification approaches on two blended emotion prediction tasks: (1) predicting the presence of emotions in a given sample, and (2) predicting the relative salience of emotions in a blend. Our results show that unimodal classifiers achieve up to 29% presence accuracy and 13% salience accuracy on the validation set, while multimodal methods yield clear improvements, with ImageBind + WavLM reaching 35% presence accuracy and HiCMAE 18% salience accuracy. On the held-out test set, the best models achieve 33% presence accuracy (VideoMAEv2 + HuBERT) and 18% salience accuracy (HiCMAE). In sum, the BLEMORE dataset provides a valuable resource to advancing research on emotion recognition systems that account for the complexity and significance of blended emotion expressions.

Tanay Agrawal, Mohammed Guermal, Michal Balazia et al.

Challenges in cross-learning involve inhomogeneous or even inadequate amount of training data and lack of resources for retraining large pretrained models. Inspired by transfer learning techniques in NLP, adapters and prefix tuning, this paper presents a new model-agnostic plugin architecture for cross-learning, called CM3T, that adapts transformer-based models to new or missing information. We introduce two adapter blocks: multi-head vision adapters for transfer learning and cross-attention adapters for multimodal learning. Training becomes substantially efficient as the backbone and other plugins do not need to be finetuned along with these additions. Comparative and ablation studies on three datasets Epic-Kitchens-100, MPIIGroupInteraction and UDIVA v0.5 show efficacy of this framework on different recording settings and tasks. With only 12.8% trainable parameters compared to the backbone to process video input and only 22.3% trainable parameters for two additional modalities, we achieve comparable and even better results than the state-of-the-art. CM3T has no specific requirements for training or pretraining and is a step towards bridging the gap between a general model and specific practical applications of video classification.

Sanya Sinha, Michal Balazia, Francois Bremond

Instructional cataract surgery videos are crucial for ophthalmologists and trainees to observe surgical details repeatedly. This paper presents a deep learning model for real-time identification of surgical instruments in these videos, using a custom dataset scraped from open-access sources. Inspired by the architecture of YOLOV9, the model employs a Programmable Gradient Information (PGI) mechanism and a novel Generally-Optimized Efficient Layer Aggregation Network (Go-ELAN) to address the information bottleneck problem, enhancing Minimum Average Precision (mAP) at higher Non-Maximum Suppression Intersection over Union (NMS IoU) scores. The Go-ELAN YOLOV9 model, evaluated against YOLO v5, v7, v8, v9 vanilla, Laptool and DETR, achieves a superior mAP of 73.74 at IoU 0.5 on a dataset of 615 images with 10 instrument classes, demonstrating the effectiveness of the proposed model.

Tanay Agrawal, Dhruv Agarwal, Michal Balazia et al.

Personality computing and affective computing have gained recent interest in many research areas. The datasets for the task generally have multiple modalities like video, audio, language and bio-signals. In this paper, we propose a flexible model for the task which exploits all available data. The task involves complex relations and to avoid using a large model for video processing specifically, we propose the use of behaviour encoding which boosts performance with minimal change to the model. Cross-attention using transformers has become popular in recent times and is utilised for fusion of different modalities. Since long term relations may exist, breaking the input into chunks is not desirable, thus the proposed model processes the entire input together. Our experiments show the importance of each of the above contributions

Neelabh Sinha, Michal Balazia, Francois Bremond

3D gaze estimation is about predicting the line of sight of a person in 3D space. Person-independent models for the same lack precision due to anatomical differences of subjects, whereas person-specific calibrated techniques add strict constraints on scalability. To overcome these issues, we propose a novel technique, Facial Landmark Heatmap Activated Multimodal Gaze Estimation (FLAME), as a way of combining eye anatomical information using eye landmark heatmaps to obtain precise gaze estimation without any person-specific calibration. Our evaluation demonstrates a competitive performance of about 10% improvement on benchmark datasets ColumbiaGaze and EYEDIAP. We also conduct an ablation study to validate our method.

Michal Balazia, S L Happy, Francois Bremond et al.

Face recognition has been widely accepted as a means of identification in applications ranging from border control to security in the banking sector. Surprisingly, while widely accepted, we still lack the understanding of uniqueness or distinctiveness of faces as biometric modality. In this work, we study the impact of factors such as image resolution, feature representation, database size, age and gender on uniqueness denoted by the Kullback-Leibler divergence between genuine and impostor distributions. Towards understanding the impact, we present experimental results on the datasets AT&T, LFW, IMDb-Face, as well as ND-TWINS, with the feature extraction algorithms VGGFace, VGG16, ResNet50, InceptionV3, MobileNet and DenseNet121, that reveal the quantitative impact of the named factors. While these are early results, our findings indicate the need for a better understanding of the concept of biometric uniqueness and its implication on face recognition.

Michal Balazia, Petr Sojka

Gait recognition from motion capture data, as a pattern classification discipline, can be improved by the use of machine learning. This paper contributes to the state-of-the-art with a statistical approach for extracting robust gait features directly from raw data by a modification of Linear Discriminant Analysis with Maximum Margin Criterion. Experiments on the CMU MoCap database show that the suggested method outperforms thirteen relevant methods based on geometric features and a method to learn the features by a combination of Principal Component Analysis and Linear Discriminant Analysis. The methods are evaluated in terms of the distribution of biometric templates in respective feature spaces expressed in a number of class separability coefficients and classification metrics. Results also indicate a high portability of learned features, that means, we can learn what aspects of walk people generally differ in and extract those as general gait features. Recognizing people without needing group-specific features is convenient as particular people might not always provide annotated learning data. As a contribution to reproducible research, our evaluation framework and database have been made publicly available. This research makes motion capture technology directly applicable for human recognition.

Michal Balazia, Petr Sojka

This work offers a design of a video surveillance system based on a soft biometric -- gait identification from MoCap data. The main focus is on two substantial issues of the video surveillance scenario: (1) the walkers do not cooperate in providing learning data to establish their identities and (2) the data are often noisy or incomplete. We show that only a few examples of human gait cycles are required to learn a projection of raw MoCap data onto a low-dimensional sub-space where the identities are well separable. Latent features learned by Maximum Margin Criterion (MMC) method discriminate better than any collection of geometric features. The MMC method is also highly robust to noisy data and works properly even with only a fraction of joints tracked. The overall workflow of the design is directly applicable for a day-to-day operation based on the available MoCap technology and algorithms for gait analysis. In the concept we introduce, a walker's identity is represented by a cluster of gait data collected at their incidents within the surveillance system: They are how they walk.

Michal Balazia, Petr Sojka

As a contribution to reproducible research, this paper presents a framework and a database to improve the development, evaluation and comparison of methods for gait recognition from motion capture (MoCap) data. The evaluation framework provides implementation details and source codes of state-of-the-art human-interpretable geometric features as well as our own approaches where gait features are learned by a modification of Fisher's Linear Discriminant Analysis with the Maximum Margin Criterion, and by a combination of Principal Component Analysis and Linear Discriminant Analysis. It includes a description and source codes of a mechanism for evaluating four class separability coefficients of feature space and four rank-based classifier performance metrics. This framework also contains a tool for learning a custom classifier and for classifying a custom query on a custom gallery. We provide an experimental database along with source codes for its extraction from the general CMU MoCap database.

Michal Balazia, Petr Sojka

MoCap-based human identification, as a pattern recognition discipline, can be optimized using a machine learning approach. Yet in some applications such as video surveillance new identities can appear on the fly and labeled data for all encountered people may not always be available. This work introduces the concept of learning walker-independent gait features directly from raw joint coordinates by a modification of the Fisher Linear Discriminant Analysis with Maximum Margin Criterion. Our new approach shows not only that these features can discriminate different people than who they are learned on, but also that the number of learning identities can be much smaller than the number of walkers encountered in the real operation.

Michal Balazia, Petr Sojka

In the field of gait recognition from motion capture data, designing human-interpretable gait features is a common practice of many fellow researchers. To refrain from ad-hoc schemes and to find maximally discriminative features we may need to explore beyond the limits of human interpretability. This paper contributes to the state-of-the-art with a machine learning approach for extracting robust gait features directly from raw joint coordinates. The features are learned by a modification of Linear Discriminant Analysis with Maximum Margin Criterion so that the identities are maximally separated and, in combination with an appropriate classifier, used for gait recognition. Experiments on the CMU MoCap database show that this method outperforms eight other relevant methods in terms of the distribution of biometric templates in respective feature spaces expressed in four class separability coefficients. Additional experiments indicate that this method is a leading concept for rank-based classifier systems.