Kai Yao

Zixian Su, Jingwei Guo, Kai Yao et al.

While recent test-time adaptations exhibit efficacy by adjusting batch normalization to narrow domain disparities, their effectiveness diminishes with realistic mini-batches due to inaccurate target estimation. As previous attempts merely introduce source statistics to mitigate this issue, the fundamental problem of inaccurate target estimation still persists, leaving the intrinsic test-time domain shifts unresolved. This paper delves into the problem of mini-batch degradation. By unraveling batch normalization, we discover that the inexact target statistics largely stem from the substantially reduced class diversity in batch. Drawing upon this insight, we introduce a straightforward tool, Test-time Exponential Moving Average (TEMA), to bridge the class diversity gap between training and testing batches. Importantly, our TEMA adaptively extends the scope of typical methods beyond the current batch to incorporate a diverse set of class information, which in turn boosts an accurate target estimation. Built upon this foundation, we further design a novel layer-wise rectification strategy to consistently promote test-time performance. Our proposed method enjoys a unique advantage as it requires neither training nor tuning parameters, offering a truly hassle-free solution. It significantly enhances model robustness against shifted domains and maintains resilience in diverse real-world scenarios with various batch sizes, achieving state-of-the-art performance on several major benchmarks. Code is available at \url{https://github.com/kiwi12138/RealisticTTA}.

Tiandi Ye, Wenyan Liu, Kai Yao et al.

Federated learning (FL) is a privacy-preserving machine learning paradigm that enables collaborative model training across multiple distributed clients without disclosing their raw data. Personalized federated learning (pFL) has gained increasing attention for its ability to address data heterogeneity. However, most existing pFL methods assume that each client's data follows a single distribution and learn one client-level personalized model for each client. This assumption often fails in practice, where a single client may possess data from multiple sources or domains, resulting in significant intra-client heterogeneity and suboptimal performance. To tackle this challenge, we propose pFedBayesPT, a fine-grained instance-wise pFL framework based on visual prompt tuning. Specifically, we formulate instance-wise prompt generation from a Bayesian perspective and model the prompt posterior as an implicit distribution to capture diverse visual semantics. We derive a variational training objective under the semi-implicit variational inference framework. Extensive experiments on benchmark datasets demonstrate that pFedBayesPT consistently outperforms existing pFL methods under both feature and label heterogeneity settings.

Kai Yao, Kaizhu Huang, Jie Sun et al.

Automatic nuclei segmentation and classification play a vital role in digital pathology. However, previous works are mostly built on data with limited diversity and small sizes, making the results questionable or misleading in actual downstream tasks. In this paper, we aim to build a reliable and robust method capable of dealing with data from the 'the clinical wild'. Specifically, we study and design a new method to simultaneously detect, segment, and classify nuclei from Haematoxylin and Eosin (H&E) stained histopathology data, and evaluate our approach using the recent largest dataset: PanNuke. We address the detection and classification of each nuclei as a novel semantic keypoint estimation problem to determine the center point of each nuclei. Next, the corresponding class-agnostic masks for nuclei center points are obtained using dynamic instance segmentation. Meanwhile, we proposed a novel Joint Pyramid Fusion Module (JPFM) to model the cross-scale dependencies, thus enhancing the local feature for better nuclei detection and classification. By decoupling two simultaneous challenging tasks and taking advantage of JPFM, our method can benefit from class-aware detection and class-agnostic segmentation, thus leading to a significant performance boost. We demonstrate the superior performance of our proposed approach for nuclei segmentation and classification across 19 different tissue types, delivering new benchmark results.

Kai Yao, Kaizhu Huang, Jie Sun et al.

We consider unsupervised cell nuclei segmentation in this paper. Exploiting the recently-proposed unpaired image-to-image translation between cell nuclei images and randomly synthetic masks, existing approaches, e.g., CycleGAN, have achieved encouraging results. However, these methods usually take a two-stage pipeline and fail to learn end-to-end in cell nuclei images. More seriously, they could lead to the lossy transformation problem, i.e., the content inconsistency between the original images and the corresponding segmentation output. To address these limitations, we propose a novel end-to-end unsupervised framework called Aligned Disentangling Generative Adversarial Network (AD-GAN). Distinctively, AD-GAN introduces representation disentanglement to separate content representation (the underling spatial structure) from style representation (the rendering of the structure). With this framework, spatial structure can be preserved explicitly, enabling a significant reduction of macro-level lossy transformation. We also propose a novel training algorithm able to align the disentangled content in the latent space to reduce micro-level lossy transformation. Evaluations on real-world 2D and 3D datasets show that AD-GAN substantially outperforms the other comparison methods and the professional software both quantitatively and qualitatively. Specifically, the proposed AD-GAN leads to significant improvement over the current best unsupervised methods by an average 17.8% relatively (w.r.t. the metric DICE) on four cell nuclei datasets. As an unsupervised method, AD-GAN even performs competitive with the best supervised models, taking a further leap towards end-to-end unsupervised nuclei segmentation.

Kai Yao, Alberto Ortiz, Francisco Bonnin-Pascual

Following the success of machine vision systems for on-line automated quality control and inspection processes, an object recognition solution is presented in this work for two different specific applications, i.e., the detection of quality control items in surgery toolboxes prepared for sterilizing in a hospital, as well as the detection of defects in vessel hulls to prevent potential structural failures. The solution has two stages. First, a feature pyramid architecture based on Single Shot MultiBox Detector (SSD) is used to improve the detection performance, and a statistical analysis based on ground truth is employed to select parameters of a range of default boxes. Second, a lightweight neural network is exploited to achieve oriented detection results using a regression method. The first stage of the proposed method is capable of detecting the small targets considered in the two scenarios. In the second stage, despite the simplicity, it is efficient to detect elongated targets while maintaining high running efficiency.

Kai Yao, Alberto Ortiz, Francisco Bonnin-Pascual

It is generally accepted that one of the critical parts of current vision algorithms based on deep learning and convolutional neural networks is the annotation of a sufficient number of images to achieve competitive performance. This is particularly difficult for semantic segmentation tasks since the annotation must be ideally generated at the pixel level. Weakly-supervised semantic segmentation aims at reducing this cost by employing simpler annotations that, hence, are easier, cheaper and quicker to produce. In this paper, we propose and assess a new weakly-supervised semantic segmentation approach making use of a novel loss function whose goal is to counteract the effects of weak annotations. To this end, this loss function comprises several terms based on partial cross-entropy losses, being one of them the Centroid Loss. This term induces a clustering of the image pixels in the object classes under consideration, whose aim is to improve the training of the segmentation network by guiding the optimization. The performance of the approach is evaluated against datasets from two different industry-related case studies: while one involves the detection of instances of a number of different object classes in the context of a quality control application, the other stems from the visual inspection domain and deals with the localization of images areas whose pixels correspond to scene surface points affected by a specific sort of defect. The detection results that are reported for both cases show that, despite the differences among them and the particular challenges, the use of weak annotations do not prevent from achieving a competitive performance level for both.