We introduce Memory Selection with Contrastive Learning (MSCL), an advanced Continual Learning (CL) approach, addressing challenges in dynamic and imbalanced environments. MSCL combines Feature-Distance Based Sample Selection (FDBS) for memory management, focusing on inter-class similarities and intra-class diversity, with a contrastive learning loss (IWL) for adaptive data representation. Our evaluations on datasets like MNIST, Cifar-100, miniImageNet, PACS, and DomainNet show that MSCL not only competes with but often surpasses existing memory-based CL methods, particularly in imbalanced scenarios, enhancing both balanced and imbalanced learning performance.
Histological images are often tainted with two or more stains to reveal their underlying structures and conditions. Blind Color Deconvolution (BCD) techniques separate colors (stains) and structural information (concentrations), which is useful for the processing, data augmentation, and classification of such images. Classical BCD methods rely on a complicated optimization procedure that has to be carried out on each image independently, i.e., they are not amortized methods. In contrast, once they have been trained, deep neural networks can be used in a fast, amortized manner on unseen inputs. Unfortunately, the lack of large databases of ground truth color and concentrations has limited the development of deep models for BCD. In this work, we propose a deep variational Bayesian BCD neural network (BCD-Net) for stain separation and concentration estimation. BCD-Net is trained by maximizing the evidence lower bound of the observed images, which does not require the use of ground truth examples of stains and concentrations. Results obtained using two multicenter databases (Camelyon-17 and a stain separation benchmark) demonstrate the effectiveness of BCD-Net in the stain separation tasks, while drastically reducing the computation time compared to classical non-amortized methods.
When dealing with high-dimensional multivariate time series classification problems, a well-known difficulty is the curse of dimensionality. In this article, we propose an original approach of transposition of multidimensional data into images to tackle the task of classification. We propose a lightweight hybrid model that take this transposed data as an input. This model contains convolutional layers as a feature extractor followed by a recurrent neural network. We apply our method to a large dataset consisting of individual patient medical records. We show that our approach allows us to significantly reduce the size of a network and increase its performance by opting for a transformation of the input data.
Deep Learning (DL) models have enabled very accurate pose estimation. However, most of the existing approaches require images of relatively high resolution, since locating body parts and joints accurately is challenging, which increases the computational cost of these approaches. To overcome this limitation in this paper we propose an active perception method for high-resolution pose estimation that enables efficiently selecting the most appropriate image region for analysis and then employing a bottom-up pose estimator on the corresponding region. This allows for significantly improving the efficiency of pose estimation by selectively analyzing in high resolution only the parts of the image that contain humans. To ensure the computational efficiency of the proposed method we propose using low-resolution heat maps extracted using the same pose estimation model in order to guide the active perception process. The proposed method is model agnostic since it can be combined with any bottom-up pose estimation model in order to enable high-resolution analysis. We have experimentally evaluated the proposed method using a well-known pose estimation model, Lightweight OpenPose, demonstrating its effectiveness on three high-resolution variants of the COCO2017 dataset.
This paper proposes a novel metric learning method for fair face attribute recognition. Datasets with imbalanced sample distributions can spoil the fair discrimination ability of deep learning models. This problem becomes more complex in multi-label classification tasks due to the variety of imbalance levels for each label. To tackle this problem, we propose Imbalance-Aware Adaptive Margin (IAAM). IAAM utilizes prior information on data imbalance with new parameters that control the relative margins between classes for each label. Specifically, it adaptively tunes margins in the metric learning loss to give larger margins for minority classes and smaller margins for majority classes. The proposed method is a flexible reformulation of previous studies for multi-label classification tasks. We performed experiments on the Celeb-A and LFWA datasets and evaluated with fair classification metrics. Our findings showed that the IAAM led to excellent performance on fair face attribute recognition both in the weakly imbalanced labels and strongly imbalanced labels.
