Win-Win by Competition: Auxiliary-Free Cloth-Changing Person Re-Identification.

Recent person Re-IDentification (ReID) systems have been challenged by changes in personnel clothing, leading to the study of Cloth-Changing person ReID (CC-ReID). Commonly used techniques involve incorporating auxiliary information (e.g., body masks, gait, skeleton, and keypoints) to accurately identify the target pedestrian. However, the effectiveness of these methods heavily relies on the quality of auxiliary information and comes at the cost of additional computational resources, ultimately increasing system complexity. This paper focuses on achieving CC-ReID by effectively leveraging the information concealed within the image. To this end, we introduce an Auxiliary-free Competitive IDentification (ACID) model. It achieves a win-win situation by enriching the identity (ID)-preserving information conveyed by the appearance and structure features while maintaining holistic efficiency. In detail, we build a hierarchical competitive strategy that progressively accumulates meticulous ID cues with discriminating feature extraction at the global, channel, and pixel levels during model inference. After mining the hierarchical discriminative clues for appearance and structure features, these enhanced ID-relevant features are crosswise integrated to reconstruct images for reducing intra-class variations. Finally, by combing with self- and cross-ID penalties, the ACID is trained under a generative adversarial learning framework to effectively minimize the distribution discrepancy between the generated data and real-world data. Experimental results on four public cloth-changing datasets (i.e., PRCC-ReID, VC-Cloth, LTCC-ReID, and Celeb-ReID) demonstrate the proposed ACID can achieve superior performance over state-of-the-art methods. The code is available soon at: https://github.com/BoomShakaY/Win-CCReID.

Towards Robust Person Re-Identification by Defending Against Universal Attackers.

Recent studies show that deep person re-identification (re-ID) models are vulnerable to adversarial examples, so it is critical to improving the robustness of re-ID models against attacks. To achieve this goal, we explore the strengths and weaknesses of existing re-ID models, i.e., designing learning-based attacks and training robust models by defending against the learned attacks. The contributions of this paper are three-fold: First, we build a holistic attack-defense framework to study the relationship between the attack and defense for person re-ID. Second, we introduce a combinatorial adversarial attack that is adaptive to unseen domains and unseen model types. It consists of distortions in pixel and color space (i.e., mimicking camera shifts). Third, we propose a novel virtual-guided meta-learning algorithm for our attack-defense system. We leverage a virtual dataset to conduct experiments under our meta-learning framework, which can explore the cross-domain constraints for enhancing the generalization of the attack and the robustness of the re-ID model. Comprehensive experiments on three large-scale re-ID benchmarks demonstrate that: 1) Our combinatorial attack is effective and highly universal in cross-model and cross-dataset scenarios; 2) Our meta-learning algorithm can be readily applied to different attack and defense approaches, which can reach consistent improvement; 3) The defense model trained on the learning-to-learn framework is robust to recent SOTA attacks that are not even used during training.

Unsupervised Foggy Scene Understanding via Self Spatial-Temporal Label Diffusion.

Understanding foggy image sequence in driving scene is critical for autonomous driving, but it remains a challenging task due to the difficulty in collecting and annotating real-world images of adverse weather. Recently, self-training strategy has been considered as a powerful solution for unsupervised domain adaptation, which iteratively adapts the model from the source domain to the target domain by generating target pseudo labels and re-training the model. However, the selection of confident pseudo labels inevitably suffers from the conflict between sparsity and accuracy, both of which will lead to suboptimal models. To tackle this problem, we exploit the characteristics of the foggy image sequence of driving scenes to densify the confident pseudo labels. Specifically, based on the two discoveries of local spatial similarity and adjacent temporal correspondence of the sequential image data, we propose a novel Target-Domain driven pseudo label Diffusion (TDo-Dif) scheme. It employs superpixels and optical flows to identify the spatial similarity and temporal correspondence, respectively, and then diffuses the confident but sparse pseudo labels within a superpixel or a temporal corresponding pair linked by the flow. Moreover, to ensure the feature similarity of the diffused pixels, we introduce local spatial similarity loss and temporal contrastive loss in the model re-training stage. Experimental results show that our TDo-Dif scheme helps the adaptive model achieve 51.92% and 53.84% mean intersection-over-union (mIoU) on two publicly available natural foggy datasets (Foggy Zurich and Foggy Driving), which exceeds the state-of-the-art unsupervised domain adaptive semantic segmentation methods. The proposed method can also be applied to non-sequential images in the target domain by considering only spatial similarity.

Weakly-Supervised Learning With Complementary Heatmap for Retinal Disease Detection.

There are many types of retinal disease, and accurately detecting these diseases is crucial for proper diagnosis. Convolutional neural networks (CNNs) typically perform well on detection tasks, and the attention module of CNNs can generate heatmaps as visual explanations of the model. However, the generated heatmap can only detect the most discriminative part, which is problematic because many object regions may exist in the region beside the heatmap in an area known as a complementary heatmap. In this study, we developed a method specifically designed multi-retinal diseases detection from fundus images with the complementary heatmap. The proposed CAM-based method is designed for 2D color images of the retina, rather than MRI images or other forms of data. Moreover, unlike other visual images for disease detection, fundus images of multiple retinal diseases have features such as distinguishable lesion region boundaries, overlapped lesion regions between diseases, and specific pathological structures (e.g. scattered blood spots) that lead to mis-classifications. Based on these considerations, we designed two new loss functions, attention-explore loss and attention-refine loss, to generate accurate heatmaps. We select both "bad" and "good" heatmaps based on the prediction score of ground truth and train them with the two loss functions. When the detection accuracy increases, the classification performance of the model is also improved. Experiments on a dataset consisting of five diseases showed that our approach improved both the detection accuracy and the classification accuracy, and the improved heatmaps were closer to the lesion regions than those of current state-of-the-art methods.

Classification of large-scale image database of various skin diseases using deep learning.

PURPOSE: The purpose of this study was to develop a deep learning-based computer-aided diagnosis system for skin disease classification using photographic images of patients. The targets are 59 skin diseases, including localized and diffuse diseases captured by photographic cameras, resulting in highly diverse images in terms of the appearance of the diseases or photographic conditions. METHODS: ResNet-18 is used as a baseline model for classification and is reinforced by metric learning to boost generalization in classification by avoiding the overfitting of the training data and increasing the reliability of CADx for dermatologists. Patient-wise classification is performed by aggregating the inference vectors of all the input patient images. RESULTS: The experiment using 70,196 images of 13,038 patients demonstrated that classification accuracy was significantly improved by both metric learning and aggregation, resulting in patient accuracies of 0.579 for Top-1, 0.793 for Top-3, and 0.863 for Top-5. The McNemar test showed that the improvements achieved by the proposed method were statistically significant. CONCLUSION: This study presents a deep learning-based classification of 59 skin diseases using multiple photographic images of a patient. The experimental results demonstrated that the proposed classification reinforced by metric learning and aggregation of multiple input images was effective in the classification of patients with diverse skin diseases and imaging conditions.

MADGAN: unsupervised medical anomaly detection GAN using multiple adjacent brain MRI slice reconstruction.

BACKGROUND: Unsupervised learning can discover various unseen abnormalities, relying on large-scale unannotated medical images of healthy subjects. Towards this, unsupervised methods reconstruct a 2D/3D single medical image to detect outliers either in the learned feature space or from high reconstruction loss. However, without considering continuity between multiple adjacent slices, they cannot directly discriminate diseases composed of the accumulation of subtle anatomical anomalies, such as Alzheimer's disease (AD). Moreover, no study has shown how unsupervised anomaly detection is associated with either disease stages, various (i.e., more than two types of) diseases, or multi-sequence magnetic resonance imaging (MRI) scans. RESULTS: We propose unsupervised medical anomaly detection generative adversarial network (MADGAN), a novel two-step method using GAN-based multiple adjacent brain MRI slice reconstruction to detect brain anomalies at different stages on multi-sequence structural MRI: (Reconstruction) Wasserstein loss with Gradient Penalty + 100 [Formula: see text] loss-trained on 3 healthy brain axial MRI slices to reconstruct the next 3 ones-reconstructs unseen healthy/abnormal scans; (Diagnosis) Average [Formula: see text] loss per scan discriminates them, comparing the ground truth/reconstructed slices. For training, we use two different datasets composed of 1133 healthy T1-weighted (T1) and 135 healthy contrast-enhanced T1 (T1c) brain MRI scans for detecting AD and brain metastases/various diseases, respectively. Our self-attention MADGAN can detect AD on T1 scans at a very early stage, mild cognitive impairment (MCI), with area under the curve (AUC) 0.727, and AD at a late stage with AUC 0.894, while detecting brain metastases on T1c scans with AUC 0.921. CONCLUSIONS: Similar to physicians' way of performing a diagnosis, using massive healthy training data, our first multiple MRI slice reconstruction approach, MADGAN, can reliably predict the next 3 slices from the previous 3 ones only for unseen healthy images. As the first unsupervised various disease diagnosis, MADGAN can reliably detect the accumulation of subtle anatomical anomalies and hyper-intense enhancing lesions, such as (especially late-stage) AD and brain metastases on multi-sequence MRI scans.

Community Detection Using Restrained Random-Walk Similarity.

In this paper, we propose a restrained random-walk similarity method for detecting the community structures of graphs. The basic premise of our method is that the starting vertices of finite-length random walks are judged to be in the same community if the walkers pass similar sets of vertices. This idea is based on our consideration that a random walker tends to move in the community including the walker's starting vertex for some time after starting the walk. Therefore, the sets of vertices passed by random walkers starting from vertices in the same community must be similar. The idea is reinforced with two conditions. First, we exclude abnormal random walks. Random walks that depart from each vertex are executed many times, and vertices that are rarely passed by the walkers are excluded from the set of vertices that the walkers may pass. Second, we forcibly restrain random walks to an appropriate length. In our method, a random walk is terminated when the walker repeatedly visits vertices that they have already passed. Experiments on real-world networks demonstrate that our method outperforms previous techniques in terms of accuracy.

How to Extract More Information With Less Burden: Fundus Image Classification and Retinal Disease Localization With Ophthalmologist Intervention.

Image classification using convolutional neural networks (CNNs) outperforms other state-of-the-art methods. Moreover, attention can be visualized as a heatmap to improve the explainability of results of a CNN. We designed a framework that can generate heatmaps reflecting lesion regions precisely. We generated initial heatmaps by using a gradient-based classification activation map (Grad-CAM). We assume that these Grad-CAM heatmaps correctly reveal the lesion regions; then we apply the attention mining technique to these heatmaps to obtain integrated heatmaps. Moreover, we assume that these Grad-CAM heatmaps incorrectly reveal the lesion regions and design a dissimilarity loss to increase their discrepancy with the Grad-CAM heatmaps. In this study, we found that having professional ophthalmologists select 30% of the heatmaps covering the lesion regions led to better results, because this step integrates (prior) clinical knowledge into the system. Furthermore, we design a knowledge preservation loss that minimizes the discrepancy between heatmaps generated from the updated CNN model and the selected heatmaps. Experiments using fundus images revealed that our method improved classification accuracy and generated attention regions closer to the ground truth lesion regions in comparison with existing methods.

Signal-to-Noise Ratio of Brillouin Grating Measurement with Micrometer-Resolution Optical Low Coherence Reflectometry.

Signal-dependent speckle-like noise was the dominant noise in a Brillouin grating measurement with micrometer-resolution optical low coherence reflectometry (OLCR). The noise was produced by the interaction of a Stokes signal with beat noise caused by a leaked pump light via square-law detection. The resultant signal-to-noise ratio (SNR) was calculated and found to be proportional to the square root of the dynamic range (DR) defined by the ratio of the Stokes signal magnitude to the variance of the beat noise. The calculation showed that even when we achieved a DR of 20 dB on a logarithmic scale, the SNR value was only 7 on a linear scale and the detected signal tended to fluctuate over ±14% with respect to the mean level. We achieved an SNR of 24 by attenuating the pump light power entering the balanced mixer by 55 dB, and this success enabled us to measure the Brillouin spectrum distributions of mated fiber connectors and a 3-dB fused fiber coupler with a micrometer resolution as examples of OLCR diagnosis.

Learning Sparse and Identity-Preserved Hidden Attributes for Person Re-Identification.

Person re-identification (Re-ID) aims at matching person images captured in non-overlapping camera views. To represent person appearance, low-level visual features are sensitive to environmental changes, while high-level semantic attributes, such as "short-hair" or "long-hair", are relatively stable. Hence, researches have started to design semantic attributes to reduce the visual ambiguity. However, to train a prediction model for semantic attributes, it requires plenty of annotations, which are hard to obtain in practical large-scale applications. To alleviate the reliance on annotation efforts, we propose to incrementally generate Deep Hidden Attribute (DHA) based on baseline deep network for newly uncovered annotations. In particular, we propose an auto-encoder model that can be plugged into any deep network to mine latent information in an unsupervised manner. To optimize the effectiveness of DHA, we reform the auto-encoder model with additional orthogonal generation module, along with identity-preserving and sparsity constraints. 1) Orthogonally generating: In order to make DHAs different from each other, Singular Vector Decomposition (SVD) is introduced to generate DHAs orthogonally. 2) Identity-preserving constraint: The generated DHAs should be distinct for telling different persons, so we associate DHAs with person identities. 3) Sparsity constraint: To enhance the discriminability of DHAs, we also introduce the sparsity constraint to restrict the number of effective DHAs for each person. Experiments conducted on public datasets have validated the effectiveness of the proposed network. On two large-scale datasets, i.e., Market-1501 and DukeMTMC-reID, the proposed method outperforms the state-of-the-art methods.

Evaluation of deep convolutional neural networks for glaucoma detection.

PURPOSE: To investigate the performance of deep convolutional neural networks (DCNNs) for glaucoma discrimination using color fundus images STUDY DESIGN: A retrospective study PATIENTS AND METHODS: To investigate the discriminative ability of 3 DCNNs, we used a total of 3312 images consisting of 369 images from glaucoma-confirmed eyes, 256 images from glaucoma-suspected eyes diagnosed by a glaucoma expert, and 2687 images judged to be nonglaucomatous eyes by a glaucoma expert. We also investigated the effects of image size on the discriminative ability and heatmap analysis to determine which parts of the image contribute to the discrimination. Additionally, we used 465 poor-quality images to investigate the effect of poor image quality on the discriminative ability. RESULTS: Three DCNNs showed areas under the curve (AUCs) of 0.9 or more. The AUC of the DCNN using glaucoma-confirmed eyes against nonglaucomatous eyes was higher than that using glaucoma-suspected eyes against nonglaucomatous eyes by approximately 0.1. The image size did not affect the discriminative ability. Heatmap analysis showed that the optic disc area was the most important area for the discrimination of glaucoma. The image quality affected the discriminative ability, and the inclusion of poor-quality images in the analysis reduced the AUC by 0.1 to 0.2. CONCLUSIONS: DCNNs may be a useful tool for detecting glaucoma or glaucoma-suspected eyes by use of fundus color images. Proper preprocessing and collection of qualified images are essential to improving the discriminative ability.

Measurement of a true Brillouin grating distribution generated at mated fiber connectors with optical low coherence reflectometry coupled with dispersive Fourier spectroscopy.

We previously reported a reflectogram from mated fiber connectors that was measured at a spatial resolution of 100 µm with Brillouin-gating-based optical low coherence reflectometry, and that agreed with a theoretical curve calculated by assuming that there was a step-like Brillouin grating distribution [Electron. Lett.53, 423 (2017)]. The agreement meant that the reflectogram was determined by the coherence function of the low coherence light and did not mean that we could observe the Brillouin grating distribution generated around the fiber connector joint. In this paper, we focused on increasing the spatial resolution to reveal the actual distribution by broadening the low coherence light, removing the erbium-doped fiber amplifier from the probe port of the interferometer to reduce the dispersion and phase fluctuations, and introducing dispersive Fourier spectroscopy to numerically and completely eliminate the residual dispersion. Although the signal-to-noise ratio (S/N) of a reflectogram obtained by a single translation of the stage was only 4, we succeeded in increasing the S/N 15-fold by averaging 200 samples acquired with repetitive measurements while maintaining a spatial resolution of 30 µm. We were able to clearly observe 320 and 430 µm wide transitions in a Brillouin grating distribution generated around the joint.

Beat noise reduction utilizing the transient acoustic-wave response of an optical fiber in Brillouin grating-based optical low coherence reflectometry.

We periodically generated a transient Brillouin grating by using continuous-wave pump light and 50 MHz pulse-wave pump light with optical low coherence reflectometry (OLCR). We extracted the Stokes light generated by the decaying part of the grating with an optical switch, and this enabled us to block the pulse-wave pump light from entering the balanced mixer, resulting in a reduction in the noise caused by the beat between the local oscillator light and the pump light. For the first time, to the best of our knowledge, we succeeded in detecting a Stokes light with an OLCR, despite the fact that the states of polarization of the probe and pump light waves were parallel, and this encouraged us to construct a polarization-independent OLCR for diagnosing optical modules and three-dimensional objects.

Person Reidentification via Discrepancy Matrix and Matrix Metric.

Person reidentification (re-id), as an important task in video surveillance and forensics applications, has been widely studied. Previous research efforts toward solving the person re-id problem have primarily focused on constructing robust vector description by exploiting appearance's characteristic, or learning discriminative distance metric by labeled vectors. Based on the cognition and identification process of human, we propose a new pattern, which transforms the feature description from characteristic vector to discrepancy matrix. In particular, in order to well identify a person, it converts the distance metric from vector metric to matrix metric, which consists of the intradiscrepancy projection and interdiscrepancy projection parts. We introduce a consistent term and a discriminative term to form the objective function. To solve it efficiently, we utilize a simple gradient-descent method under the alternating optimization process with respect to the two projections. Experimental results on public datasets demonstrate the effectiveness of the proposed pattern as compared with the state-of-the-art approaches.

Instance Search Retrospective with Focus on TRECVID.

This paper presents an overview of the Video Instance Search benchmark which was run over a period of 6 years (2010-2015) as part of the TREC Video Retrieval (TRECVID) workshop series. The main contributions of the paper include i) an examination of the evolving design of the evaluation framework and its components (system tasks, data, measures); ii) an analysis of the influence of topic characteristics (such as rigid/non rigid, planar/non-planar, stationary/mobile on performance; iii) a high-level overview of results and best-performing approaches. The Instance Search (INS) benchmark worked with a variety of large collections of data including Sound & Vision, Flickr, BBC (British Broadcasting Corporation) Rushes for the first 3 pilot years and with the small world of the BBC Eastenders series for the last 3 years.

Web Image Search Re-Ranking With Click-Based Similarity and Typicality.

In image search re-ranking, besides the well-known semantic gap, intent gap, which is the gap between the representation of users' query/demand and the real intent of the users, is becoming a major problem restricting the development of image retrieval. To reduce human effects, in this paper, we use image click-through data, which can be viewed as the implicit feedback from users, to help overcome the intention gap, and further improve the image search performance. Generally, the hypothesis-visually similar images should be close in a ranking list-and the strategy-images with higher relevance should be ranked higher than others-are widely accepted. To obtain satisfying search results, thus, image similarity and the level of relevance typicality are determinate factors correspondingly. However, when measuring image similarity and typicality, conventional re-ranking approaches only consider visual information and initial ranks of images, while overlooking the influence of click-through data. This paper presents a novel re-ranking approach, named spectral clustering re-ranking with click-based similarity and typicality. First, to learn an appropriate similarity measurement, we propose click-based multi-feature similarity learning algorithm, which conducts metric learning based on click-based triplets selection, and integrates multiple features into a unified similarity space via multiple kernel learning. Then, based on the learnt click-based image similarity measure, we conduct spectral clustering to group visually and semantically similar images into same clusters, and get the final re-rank list by calculating click-based clusters typicality and within-clusters click-based image typicality in descending order. Our experiments conducted on two real-world query-image data sets with diverse representative queries show that our proposed re-ranking approach can significantly improve initial search results, and outperform several existing re-ranking approaches.

When face-tracking meets social networks: a story of politics in news videos.

In the age of data processing, news videos are rich mines of information. After all, the news are essentially created to convey information to the public. But can we go beyond what is directly presented to us and see a wider picture? Many works already focus on what we can discover and understand from the analysis of years of news broadcasting. These analysis bring monitoring and understanding of the activity of public figures, political strategies, explanation and even prediction of critical media events. Such tools can help public figures in managing their public image, as well as support the work of journalists, social scientists and other media experts. News analysis can also be seen from the lens of complex systems, gathering many types of entities, attributes and interactions over time. As many public figures intervene in different news stories, a first interesting task is to observe the social interactions between these actors. Towards this goal, we propose to use video analysis to automatise the process of constructing social networks directly from news video archives. In this paper we are introducing a system deriving multiple social networks from face detections in news videos. We present preliminary results obtained from analysis of these networks, by monitoring the activity of more than a hundred public figures. We finally use these networks as a support for political studies and we provide an overview of the political landscape presented by the Japanese public broadcaster NHK over a decade of the 7 PM news archives.

Pleiotropic effects of the rho-kinase inhibitor fasudil after subarachnoid hemorrhage: a review of preclinical and clinical studies.

There is growing evidence that Rho-kinase contributes to cardiovascular disease, which has made Rho-kinase a target for the treatment of human diseases. To date, the only Rho-kinase inhibitor employed clinically in humans is fasudil, which has been used for the prevention of cerebral vasospasm and subsequent ischemic injury after surgery for subarachnoid hemorrhage (SAH). A number of pathological processes, in particular hemodynamic dysfunctions and inflammatory reactions, are thought to be related in the pathogenesis of delayed cerebral vasospasm and subsequent ischemic injury after SAH. This review focuses on fasudil's pleiotropic therapeutic effects: amelioration of hemodynamic dysfunction and inflammation, and discusses in detail the clinical studies on fasudil administered after the occurrence of SAH.

Evidence of a direct cellular protective effect of Rho-kinase inhibitors on endothelin-induced cardiac myocyte hypertrophy.

Using a cellular approach, the present study examined whether fasudil and active metabolite hydroxyfasudil, Rho-kinase inhibitors, exert a direct protective effect on endothelin-induced cardiac myocyte hypertrophy in vitro. Treatment with endothelin (10nM) caused significant hypertrophy of cultured neonatal rat cardiomyocytes by a 21.2% increase in cell surface area. Fasudil (1-10 µM) and hydroxyfasudil (0.3-10 µM) significantly prevented endothelin-induced cardiomyocyte hypertrophy. The present results suggest that inhibition of cardiac hypertrophy by fasudil is, at least in part, due to direct protection of cardiomyocytes from hypertrophy.

Antivasospastic effects of hydroxyfasudil, a Rho-kinase inhibitor, after subarachnoid hemorrhage.

We investigated the anti-vasospastic potential of fasudil's active metabolite, hydroxyfasudil, a Rho-kinase inhibitor, after subarachnoid hemorrhage (SAH) and also its effect on hemorheological abnormalities following cerebral ischemia. Chronic cerebral vasospasm was produced using a two-hemorrhage canine model. On day 7, angiographic vasospasm was observed in all animals, and intravenous administration of hydroxyfasudil (3 mg·kg(-1)·30 min(-1)) significantly reversed the vasospasm (predose diameter of the basilar artery, 57.9% ± 2.0% of the baseline before the injection of blood; postdose diameter, 64.5% ± 1.9%). The viscosity of whole blood was significantly increased 24 h after 1 h middle cerebral artery occlusion in rats. Hydroxyfasudil (3 and 10 mg/kg, i.p.) significantly decreased blood viscosity. The specificity of hydroxyfasudil was examined against a panel of 17 protein kinases using ELISA analysis. Hydroxyfasudil inhibited Rho-kinase α and ß at a concentration of 10 µM by 97.6% and 97.7%, respectively. No other protein kinase was inhibited with 10 µM hydroxyfasudil by over 40%. The present results indicate hydroxyfasudil is a selective inhibitor of Rho-kinase. The results also suggest that hydroxyfasudil contributes to the potency of fasudil to prevent cerebral vasospasm and hyperviscosity and suggest the potential utility of hydroxyfasudil as a therapeutic agent for patients with SAH.