Hayaru Shouno

Dewen Zhang, Wangpeng An, Hayaru Shouno

Current vision-language multimodal models are well-adapted for general visual understanding tasks. However, they perform inadequately when handling complex visual tasks related to human poses and actions due to the lack of specialized vision-language instruction-following data. We introduce a method for generating such data by integrating human keypoints with traditional visual features such as captions and bounding boxes, enabling more precise understanding of human-centric scenes. Our approach constructs a dataset comprising 200,328 samples tailored to fine-tune models for human-centric tasks, focusing on three areas: conversation, detailed description, and complex reasoning. We establish a benchmark called Human Pose and Action Understanding Benchmark (HPAUB) to assess model performance on human pose and action understanding. We fine-tune the LLaVA-1.5-7B model using this dataset and evaluate it on the benchmark, achieving significant improvements. Experimental results show an overall improvement of 21.18% compared to the original LLaVA-1.5-7B model. These findings highlight the effectiveness of keypoint-integrated data in enhancing multimodal models. Code is available at https://github.com/Ody-trek/Keypoint-Instruction-Tuning.

Dewen Zhang, Tahir Hussain, Wangpeng An et al.

Human pose estimation traditionally relies on architectures that encode keypoint priors, limiting their generalization to novel poses or unseen keypoints. Recent language-guided approaches like LocLLM reformulate keypoint localization as a vision-language task, enabling zero-shot generalization through textual descriptions. However, LocLLM's linear projector fails to capture complex spatial-textual interactions critical for high-precision localization. To address this, we propose PoseLLM, the first Large Language Model (LLM)-based pose estimation framework that replaces the linear projector with a nonlinear MLP vision-language connector. This lightweight two-layer MLP with GELU activation enables hierarchical cross-modal feature transformation, enhancing the fusion of visual patches and textual keypoint descriptions. Trained exclusively on COCO data, PoseLLM achieves 77.8 AP on the COCO validation set, outperforming LocLLM by +0.4 AP, while maintaining strong zero-shot generalization on Human-Art and MPII. Our work demonstrates that a simple yet powerful nonlinear connector significantly boosts localization accuracy without sacrificing generalization, advancing the state-of-the-art in language-guided pose estimation. Code is available at https://github.com/Ody-trek/PoseLLM.

Dewen Zhang, Tahir Hussain, Wangpeng An et al.

Current vision-language models (VLMs) are well-adapted for general visual understanding tasks. However, they perform inadequately when handling complex visual tasks related to human poses and actions due to the lack of specialized vision-language instruction-following data. We introduce a method for generating such data by integrating human keypoints with traditional visual features such as captions and bounding boxes, enabling more precise understanding of human-centric scenes. Our approach constructs a dataset comprising 200,328 samples tailored to fine-tune models for human-centric tasks, focusing on three areas: conversation, detailed description, and complex reasoning. We establish an Extended Human Pose and Action Understanding Benchmark (E-HPAUB) to assess model performance on human pose and action understanding. We fine-tune the LLaVA-1.5-7B model using this dataset and evaluate our resulting LLaVA-Pose model on the benchmark, achieving significant improvements. Experimental results show an overall improvement of 33.2% compared to the original LLaVA-1.5-7B model. These findings highlight the effectiveness of keypoint-integrated data in enhancing multimodal models for human-centric visual understanding. Code is available at https://github.com/Ody-trek/LLaVA-Pose.

Yuta Sumiya, Hayaru shouno

Machine-learning models, which are known to accurately predict patterns from large datasets, are crucial in decision making. Consequently, counterfactual explanations-methods explaining predictions by introducing input perturbations-have become prominent. These perturbations often suggest ways to alter the predictions, leading to actionable recommendations. However, the current techniques require resolving the optimization problems for each input change, rendering them computationally expensive. In addition, traditional encoding methods inadequately address the perturbations of categorical variables in tabular data. Thus, this study propose FastDCFlow, an efficient counterfactual explanation method using normalizing flows. The proposed method captures complex data distributions, learns meaningful latent spaces that retain proximity, and improves predictions. For categorical variables, we employed TargetEncoding, which respects ordinal relationships and includes perturbation costs. The proposed method outperformed existing methods in multiple metrics, striking a balance between trade offs for counterfactual explanations. The source code is available in the following repository: https://github.com/sumugit/FastDCFlow.

Sho Ichigozaki, Takahiro Kawashima, Hayaru Shouno

In this paper, we propose a Bayesian Graphical LASSO for correlated countable data and apply it to spatial crime data. In the proposed model, we assume a Gaussian Graphical Model for the latent variables which dominate the potential risks of crimes. To evaluate the proposed model, we determine optimal hyperparameters which represent samples better. We apply the proposed model for estimation of the sparse inverse covariance of the latent variable and evaluate the partial correlation coefficients. Finally, we illustrate the results on crime spots data and consider the estimated latent variables and the partial correlation coefficients of the sparse inverse covariance.

Genta Kobayashi, Hayaru Shouno

One of the methods used in image recognition is the Deep Convolutional Neural Network (DCNN). DCNN is a model in which the expressive power of features is greatly improved by deepening the hidden layer of CNN. The architecture of CNNs is determined based on a model of the visual cortex of mammals. There is a model called Residual Network (ResNet) that has a skip connection. ResNet is an advanced model in terms of the learning method, but it has not been interpreted from a biological viewpoint. In this research, we investigate the receptive fields of a ResNet on the classification task in ImageNet. We find that ResNet has orientation selective neurons and double opponent color neurons. In addition, we suggest that some inactive neurons in the first layer of ResNet affect the classification task.

Takahiro Kawashima, Hayaru Shouno

Photon-limited images are often seen in fields such as medical imaging. Although the number of collected photons on an image sensor statistically follows Poisson distribution, this type of noise is intractable, unlike Gaussian noise. In this study, we propose a Bayesian restoration method of Poisson corrupted image using Integrated Nested Laplace Approximation (INLA), which is a computational method to evaluate marginalized posterior distributions of latent Gaussian models (LGMs). When the original image can be regarded as ICAR (intrinsic conditional auto-regressive) model reasonably, our method performs very faster than well-known ones such as loopy belief propagation-based method and Markov chain Monte Carlo (MCMC) without decreasing the accuracy.

Aiga Suzuki, Hayaru Shouno

Modeling of textures in natural images is an important task to make a microscopic model of natural images. Portilla and Simoncelli proposed a generative texture model, which is based on the mechanism of visual systems in brains, with a set of texture features and a feature matching. On the other hand, the texture features, used in Portillas' model, have redundancy between its components came from typical natural textures. In this paper, we propose a contracted texture model which provides a dimension reduction for the Portillas' feature. This model is based on a hierarchical principal components analysis using known group structure of the feature. In the experiment, we reveal effective dimensions to describe texture is fewer than the original description. Moreover, we also demonstrate how well the textures can be synthesized from the contracted texture representations.

Aiga Suzuki, Hidenori Sakanashi, Shoji Kido et al.

Transfer learning is a machine learning technique designed to improve generalization performance by using pre-trained parameters obtained from other learning tasks. For image recognition tasks, many previous studies have reported that, when transfer learning is applied to deep neural networks, performance improves, despite having limited training data. This paper proposes a two-stage feature transfer learning method focusing on the recognition of textural medical images. During the proposed method, a model is successively trained with massive amounts of natural images, some textural images, and the target images. We applied this method to the classification task of textural X-ray computed tomography images of diffuse lung diseases. In our experiment, the two-stage feature transfer achieves the best performance compared to a from-scratch learning and a conventional single-stage feature transfer. We also investigated the robustness of the target dataset, based on size. Two-stage feature transfer shows better robustness than the other two learning methods. Moreover, we analyzed the feature representations obtained from DLDs imagery inputs for each feature transfer models using a visualization method. We showed that the two-stage feature transfer obtains both edge and textural features of DLDs, which does not occur in conventional single-stage feature transfer models.

Hideo Terada, Hayaru Shouno

We are trying to implement deep neural networks in the edge computing environment for real-world applications such as the IoT(Internet of Things), the FinTech etc., for the purpose of utilizing the significant achievement of Deep Learning in recent years. Especially, we now focus algorithm implementation on FPGA, because FPGA is one of the promising devices for low-cost and low-power implementation of the edge computer. In this work, we introduce Binary-DCGAN(B-DCGAN) - Deep Convolutional GAN model with binary weights and activations, and with using integer-valued operations in forward pass(train-time and run-time). And we show how to implement B-DCGAN on FPGA(Xilinx Zynq). Using the B-DCGAN, we do feasibility study of FPGA's characteristic and performance for Deep Learning. Because the binarization and using integer-valued operation reduce the memory capacity and the number of the circuit gates, it is very effective for FPGA implementation. On the other hand, the quality of generated data from the model will be decreased by these reductions. So we investigate the influence of these reductions.

Kazuyuki Hara, Daisuke Saitoh, Hayaru Shouno

Deep learning is the state-of-the-art in fields such as visual object recognition and speech recognition. This learning uses a large number of layers, huge number of units, and connections. Therefore, overfitting is a serious problem. To avoid this problem, dropout learning is proposed. Dropout learning neglects some inputs and hidden units in the learning process with a probability, p, and then, the neglected inputs and hidden units are combined with the learned network to express the final output. We find that the process of combining the neglected hidden units with the learned network can be regarded as ensemble learning, so we analyze dropout learning from this point of view.

Hiroaki Sasaki, Michael U. Gutmann, Hayaru Shouno et al.

The statistical dependencies which independent component analysis (ICA) cannot remove often provide rich information beyond the linear independent components. It would thus be very useful to estimate the dependency structure from data. While such models have been proposed, they usually concentrated on higher-order correlations such as energy (square) correlations. Yet, linear correlations are a most fundamental and informative form of dependency in many real data sets. Linear correlations are usually completely removed by ICA and related methods, so they can only be analyzed by developing new methods which explicitly allow for linearly correlated components. In this paper, we propose a probabilistic model of linear non-Gaussian components which are allowed to have both linear and energy correlations. The precision matrix of the linear components is assumed to be randomly generated by a higher-order process and explicitly parametrized by a parameter matrix. The estimation of the parameter matrix is shown to be particularly simple because using score matching, the objective function is a quadratic form. Using simulations with artificial data, we demonstrate that the proposed method improves identifiability of non-Gaussian components by simultaneously learning their correlation structure. Applications on simulated complex cells with natural image input, as well as spectrograms of natural audio data show that the method finds new kinds of dependencies between the components.

Hayaru Shouno

We treat an image restoration problem with a Poisson noise chan- nel using a Bayesian framework. The Poisson randomness might be appeared in observation of low contrast object in the field of imaging. The noise observation is often hard to treat in a theo- retical analysis. In our formulation, we interpret the observation through the Poisson noise channel as a likelihood, and evaluate the bound of it with a Gaussian function using a latent variable method. We then introduce a Gaussian Markov random field (GMRF) as the prior for the Bayesian approach, and derive the posterior as a Gaussian distribution. The latent parameters in the likelihood and the hyperparameter in the GMRF prior could be treated as hid- den parameters, so that, we propose an algorithm to infer them in the expectation maximization (EM) framework using loopy belief propagation(LBP). We confirm the ability of our algorithm in the computer simulation, and compare it with the results of other im- age restoration frameworks.