Yoichi Sato

Liangyang Ouyang, Ruicong Liu, Yifei Huang et al.

Delving into the realm of egocentric vision, the advancement of referring video object segmentation (RVOS) stands as pivotal in understanding human activities. However, existing RVOS task primarily relies on static attributes such as object names to segment target objects, posing challenges in distinguishing target objects from background objects and in identifying objects undergoing state changes. To address these problems, this work proposes a novel action-aware RVOS setting called ActionVOS, aiming at segmenting only active objects in egocentric videos using human actions as a key language prompt. This is because human actions precisely describe the behavior of humans, thereby helping to identify the objects truly involved in the interaction and to understand possible state changes. We also build a method tailored to work under this specific setting. Specifically, we develop an action-aware labeling module with an efficient action-guided focal loss. Such designs enable ActionVOS model to prioritize active objects with existing readily-available annotations. Experimental results on VISOR dataset reveal that ActionVOS significantly reduces the mis-segmentation of inactive objects, confirming that actions help the ActionVOS model understand objects' involvement. Further evaluations on VOST and VSCOS datasets show that the novel ActionVOS setting enhances segmentation performance when encountering challenging circumstances involving object state changes. We will make our implementation available at https://github.com/ut-vision/ActionVOS.

Quan Kong, Yuki Kawana, Rajat Saini et al.

In this paper, we address the challenge of fine-grained video event understanding in traffic scenarios, vital for autonomous driving and safety. Traditional datasets focus on driver or vehicle behavior, often neglecting pedestrian perspectives. To fill this gap, we introduce the WTS dataset, highlighting detailed behaviors of both vehicles and pedestrians across over 1.2k video events in hundreds of traffic scenarios. WTS integrates diverse perspectives from vehicle ego and fixed overhead cameras in a vehicle-infrastructure cooperative environment, enriched with comprehensive textual descriptions and unique 3D Gaze data for a synchronized 2D/3D view, focusing on pedestrian analysis. We also pro-vide annotations for 5k publicly sourced pedestrian-related traffic videos. Additionally, we introduce LLMScorer, an LLM-based evaluation metric to align inference captions with ground truth. Using WTS, we establish a benchmark for dense video-to-text tasks, exploring state-of-the-art Vision-Language Models with an instance-aware VideoLLM method as a baseline. WTS aims to advance fine-grained video event understanding, enhancing traffic safety and autonomous driving development.

Nie Lin, Takehiko Ohkawa, Mingfang Zhang et al.

We present a contrastive learning framework based on in-the-wild hand images tailored for pre-training 3D hand pose estimators, dubbed HandCLR. Pre-training on large-scale images achieves promising results in various tasks, but prior 3D hand pose pre-training methods have not fully utilized the potential of diverse hand images accessible from in-the-wild videos. To facilitate scalable pre-training, we first prepare an extensive pool of hand images from in-the-wild videos and design our method with contrastive learning. Specifically, we collected over 2.0M hand images from recent human-centric videos, such as 100DOH and Ego4D. To extract discriminative information from these images, we focus on the similarity of hands; pairs of similar hand poses originating from different samples, and propose a novel contrastive learning method that embeds similar hand pairs closer in the latent space. Our experiments demonstrate that our method outperforms conventional contrastive learning approaches that produce positive pairs sorely from a single image with data augmentation. We achieve significant improvements over the state-of-the-art method in various datasets, with gains of 15% on FreiHand, 10% on DexYCB, and 4% on AssemblyHands.

Hiroyuki Chuma, Kanji Otsuka, Yoichi Sato

This paper reports an unexpected finding: in a deterministic hyperdimensional computing (HDC) architecture **that inverts the conventional role of Galois-field algebra -- employing it not for error correction toward a unique answer but as an engine for relative similarity and path-quality ranking -- **a path-dependent semantic selection mechanism emerges, equivalent to spike-timing-dependent plasticity (STDP), with magnitude predictable a priori from a closed-form expression matching measured values. Addressing catastrophic forgetting, learning stagnation, and the Binding Problem at an algebraic level, we propose VaCoAl (Vague Coincident Algorithm) and its Python implementation PyVaCoAl on ultra-high-dimensional SRAM/DRAM-CAM. Rooted in Sparse Distributed Memory, it resolves orthogonalisation and retrieval in high-dimensional binary spaces via Galois-field diffusion, enabling low-load deployment. Crucially, VaCoAl embeds a cognitive bound -- the Frontier Size -- into its architecture, ranking candidates by path-integral confidence (CR2) to achieve compositional generalisation; this bounded-rationality design produces STDP-like selection that error-correction paradigms structurally cannot attain. We evaluated multi-hop reasoning on about 470k mentor-student relations from Wikidata, tracing up to 57 generations (over 25.5M paths). HDC bundling and unbinding with CR-based denoising quantify concept propagation over DAGs. Results show a reinterpretation of the Newton-Leibniz dispute and a phase transition from sparse convergence to a post-Leibniz "superhighway", with structural indicators supporting a Kuhnian paradigm shift. VaCoAl thus defines a third paradigm, HDC-AI, complementing LLMs with reversible, auditable multi-hop reasoning.

Hiroyuki Chuma, Kanji Otsuka, Yoichi Sato

Vector-HaSH and the Tolman-Eichenbaum Machine (TEM) propose that the hippocampal-entorhinal circuit factorizes content from a prestructured grid-cell scaffold and supports compositional memory via ripple-mediated replay. Human iEEG shows that hippocampal sharp-wave ripples (SWRs) gate episodic recall, ripple-locked cortical reactivation recapitulates encoding-time patterns, and multi-hop replay fidelity decays multiplicatively along sequence length. These literatures have advanced in parallel without a shared algebraic object. We show that VaCoAl, an algebro-deterministic hyperdimensional memory architecture built from Galois-field LFSRs, supplies that object. Specifically, deterministic Galois-field diffusion provides a substrate-level alternative to Vector-HaSH's random scaffold-to-hippocampus projection that satisfies the same quasi-orthogonality requirement, with matched second-moment statistics, stronger avalanche behavior, and bit-exact reproducibility. The path-integral Confidence Ratio $CR_2$, the product of per-step $CR_1$ values along an $n$-hop chain, is the natural functional form for multi-hop replay-fidelity decay under conditional independence of per-step reactivation, providing the first algebraically tractable model of reported multiplicative decay. STDP-like path selection in VaCoAl follows from architectural demands -- similarity preservation, compositional reversibility, and bounded-frontier search -- that also constrain hippocampal computation. We further argue that VaCoAl operating regimes share architectural commitments with the EC--CA3 and EC--DG--CA3 pathways, motivating an energy-capacity-plasticity reading of why both are conserved across $>$520 Myr of evolution and primate dentate-gyrus elaboration. We prove formal correspondences, derive testable iEEG predictions, and bridge computational neuroscience and hyperdimensional engineering.

Hiroyuki Chuma, Kanji Otsuk, Yoichi Sato

In The Algebraic Mind, Gary Marcus identified three components essential for any adequate cognitive architecture: operations over variables, recursively structured representations, and a distinction between mental representations of individuals and kinds. He argued that standard multilayer perceptrons supported none of these, acknowledging that a neural implementation using registers and treelets, constructed via developmental programs rather than gradient descent, remained a programmatic conjecture. Twenty-five years later, the required substrate is now available. Our newly developed PyVaCoAl/VaCoAl is a hyperdimensional computing architecture organized end-to-end around a single algebraic primitive: XOR-and-shift over GF(2), implemented by primitive-polynomial linear-feedback shift registers. The architecture supports reversible variable binding via Bind(R,F) = R XOR shift(F), non-commutative compositional bundling that distinguishes "the dog bites the man" from "the man bites the dog," and address-space individual/kind separation under the same algebra. A companion perspective argues that the dentate gyrus-CA3 circuit is a biological homologue of this same engine, with developmentally specified mossy-fiber targeting supplying the innate microcircuitry Marcus anticipated. In this paper, we map the correspondence between Marcus's three pillars and the operational commitments of PyVaCoAl/VaCoAl. We reinterpret the treelet as an algebraic register set indexed by a primitive generator polynomial, arguing that this architecture provides a functional neural substrate meeting Marcus's specifications far more closely than the tensor products, circular convolution, or temporal synchrony available in 2001. We also demonstrate how this substrate naturally extends to Pearl's rung-3 counterfactual reasoning, a capability the original treelet program did not directly target.

Nie Lin, Takehiko Ohkawa, Yifei Huang et al.

We present a framework for pre-training of 3D hand pose estimation from in-the-wild hand images sharing with similar hand characteristics, dubbed SimHand. Pre-training with large-scale images achieves promising results in various tasks, but prior methods for 3D hand pose pre-training have not fully utilized the potential of diverse hand images accessible from in-the-wild videos. To facilitate scalable pre-training, we first prepare an extensive pool of hand images from in-the-wild videos and design our pre-training method with contrastive learning. Specifically, we collect over 2.0M hand images from recent human-centric videos, such as 100DOH and Ego4D. To extract discriminative information from these images, we focus on the similarity of hands: pairs of non-identical samples with similar hand poses. We then propose a novel contrastive learning method that embeds similar hand pairs closer in the feature space. Our method not only learns from similar samples but also adaptively weights the contrastive learning loss based on inter-sample distance, leading to additional performance gains. Our experiments demonstrate that our method outperforms conventional contrastive learning approaches that produce positive pairs sorely from a single image with data augmentation. We achieve significant improvements over the state-of-the-art method (PeCLR) in various datasets, with gains of 15% on FreiHand, 10% on DexYCB, and 4% on AssemblyHands. Our code is available at https://github.com/ut-vision/SiMHand.

Yifei Huang, Xiaoxiao Li, Lijin Yang et al.

The human gaze is a cost-efficient physiological data that reveals human underlying attentional patterns. The selective attention mechanism helps the cognition system focus on task-relevant visual clues by ignoring the presence of distractors. Thanks to this ability, human beings can efficiently learn from a very limited number of training samples. Inspired by this mechanism, we aim to leverage gaze for medical image analysis tasks with small training data. Our proposed framework includes a backbone encoder and a Selective Attention Network (SAN) that simulates the underlying attention. The SAN implicitly encodes information such as suspicious regions that is relevant to the medical diagnose tasks by estimating the actual human gaze. Then we design a novel Auxiliary Attention Block (AAB) to allow information from SAN to be utilized by the backbone encoder to focus on selective areas. Specifically, this block uses a modified version of a multi-head attention layer to simulate the human visual search procedure. Note that the SAN and AAB can be plugged into different backbones, and the framework can be used for multiple medical image analysis tasks when equipped with task-specific heads. Our method is demonstrated to achieve superior performance on both 3D tumor segmentation and 2D chest X-ray classification tasks. We also show that the estimated gaze probability map of the SAN is consistent with an actual gaze fixation map obtained by board-certified doctors.

Lijin Yang, Yifei Huang, Yusuke Sugano et al.

In this report, we describe the technical details of our submission to the 2021 EPIC-KITCHENS-100 Unsupervised Domain Adaptation Challenge for Action Recognition. Leveraging multiple modalities has been proved to benefit the Unsupervised Domain Adaptation (UDA) task. In this work, we present Multi-Modal Mutual Enhancement Module (M3EM), a deep module for jointly considering information from multiple modalities to find the most transferable representations across domains. We achieve this by implementing two sub-modules for enhancing each modality using the context of other modalities. The first sub-module exchanges information across modalities through the semantic space, while the second sub-module finds the most transferable spatial region based on the consensus of all modalities.

Yoichi Sato, Yasuhiko Takegami, Takamune Asamoto et al.

[Objective] To develop a Computer-aided diagnosis (CAD) system for plane frontal hip X-rays with a deep learning model trained on a large dataset collected at multiple centers. [Materials and Methods]. We included 5295 cases with neck fracture or trochanteric fracture who were diagnosed and treated by orthopedic surgeons using plane X-rays or computed tomography (CT) or magnetic resonance imaging (MRI) who visited each institution between April 2009 and March 2019 were enrolled. Cases in which both hips were not included in the photographing range, femoral shaft fractures, and periprosthetic fractures were excluded, and 5242 plane frontal pelvic X-rays obtained from 4,851 cases were used for machine learning. These images were divided into 5242 images including the fracture side and 5242 images without the fracture side, and a total of 10484 images were used for machine learning. A deep convolutional neural network approach was used for machine learning. Pytorch 1.3 and Fast.ai 1.0 were used as frameworks, and EfficientNet-B4, which is pre-trained ImageNet model, was used. In the final evaluation, accuracy, sensitivity, specificity, F-value and area under the curve (AUC) were evaluated. Gradient-weighted class activation mapping (Grad-CAM) was used to conceptualize the diagnostic basis of the CAD system. [Results] The diagnostic accuracy of the learning model was accuracy of 96. 1 %, sensitivity of 95.2 %, specificity of 96.9 %, F-value of 0.961, and AUC of 0.99. The cases who were correct for the diagnosis showed generally correct diagnostic basis using Grad-CAM. [Conclusions] The CAD system using deep learning model which we developed was able to diagnose hip fracture in the plane X-ray with the high accuracy, and it was possible to present the decision reason.

Minjie Cai, Kris Kitani, Yoichi Sato

This paper proposes a novel method for understanding daily hand-object manipulation by developing computer vision-based techniques. Specifically, we focus on recognizing hand grasp types, object attributes and manipulation actions within an unified framework by exploring their contextual relationships. Our hypothesis is that it is necessary to jointly model hands, objects and actions in order to accurately recognize multiple tasks that are correlated to each other in hand-object manipulation. In the proposed model, we explore various semantic relationships between actions, grasp types and object attributes, and show how the context can be used to boost the recognition of each component. We also explore the spatial relationship between the hand and object in order to detect the manipulated object from hand in cluttered environment. Experiment results on all three recognition tasks show that our proposed method outperforms traditional appearance-based methods which are not designed to take into account contextual relationships involved in hand-object manipulation. The visualization and generalizability study of the learned context further supports our hypothesis.