Image classification has been one of the most popular tasks in Deep Learning, seeing an abundance of impressive implementations each year. However, there is a lot of criticism tied to promoting complex architectures that continuously push performance metrics higher and higher. Robustness tests can uncover several vulnerabilities and biases which go unnoticed during the typical model evaluation stage. So far, model robustness under distribution shifts has mainly been examined within carefully curated datasets. Nevertheless, such approaches do not test the real response of classifiers in the wild, e.g. when uncurated web-crawled image data of corresponding classes are provided. In our work, we perform fine-grained classification on closely related categories, which are identified with the help of hierarchical knowledge. Extensive experimentation on a variety of convolutional and transformer-based architectures reveals model robustness in this novel setting. Finally, hierarchical knowledge is again employed to evaluate and explain misclassifications, providing an information-rich evaluation scheme adaptable to any classifier.
Graves' disease is a common condition that is diagnosed clinically by determining the smoothness of the thyroid texture and its morphology in ultrasound images. Currently, the most widely used approach for the automated diagnosis of Graves' disease utilizes Convolutional Neural Networks (CNNs) for both feature extraction and classification. However, these methods demonstrate limited efficacy in capturing texture features. Given the high capacity of wavelets in describing texture features, this research integrates learnable wavelet modules utilizing the Lifting Scheme into CNNs and incorporates a parallel wavelet branch into the ResNet18 model to enhance texture feature extraction. Our model can analyze texture features in spatial and frequency domains simultaneously, leading to optimized classification accuracy. We conducted experiments on collected ultrasound datasets and publicly available natural image texture datasets, our proposed network achieved 97.27% accuracy and 95.60% recall on ultrasound datasets, 60.765% accuracy on natural image texture datasets, surpassing the accuracy of ResNet and conrming the effectiveness of our approach.
The Computer_Aided Diagnosis (CAD) systems facilitate accurate diagnosis of diseases. The development of CADs by leveraging third generation neural network, namely, Spiking Neural Network (SNN), is essential to utilize of the benefits of SNNs, such as their event_driven processing, parallelism, low power consumption, and the ability to process sparse temporal_spatial information. However, Deep SNN as a deep learning model faces challenges with unreliability. To deal with unreliability challenges due to inability to quantify the uncertainty of the predictions, we proposed a deep Bayesian Convolutional Spiking Neural Network based_CADs with uncertainty_aware module. In this study, the Monte Carlo Dropout method as Bayesian approximation is used as an uncertainty quantification method. This method was applied to several medical image classification tasks. Our experimental results demonstrate that our proposed model is accurate and reliable and will be a proper alternative to conventional deep learning for medical image classification.
Due to the limited amount and imbalanced classes of labeled training data, the conventional supervised learning can not ensure the discrimination of the learned feature for hyperspectral image (HSI) classification. In this paper, we propose a modified diversity of class probability estimation (MDCPE) with two deep neural networks to learn spectral-spatial feature for HSI classification. In co-training phase, recurrent neural network (RNN) and convolutional neural network (CNN) are utilized as two learners to extract features from labeled and unlabeled data. Based on the extracted features, MDCPE selects most credible samples to update initial labeled data by combining k-means clustering with the traditional diversity of class probability estimation (DCPE) co-training. In this way, MDCPE can keep new labeled data class-balanced and extract discriminative features for both the minority and majority classes. During testing process, classification results are acquired by co-decision of the two learners. Experimental results demonstrate that the proposed semi-supervised co-training method can make full use of unlabeled information to enhance generality of the learners and achieve favorable accuracies on all three widely used data sets: Salinas, Pavia University and Pavia Center.
Conditional Generative Adversarial Networks (cGANs) extend the standard unconditional GAN framework to learning joint data-label distributions from samples, and have been established as powerful generative models capable of generating high-fidelity imagery. A challenge of training such a model lies in properly infusing class information into its generator and discriminator. For the discriminator, class conditioning can be achieved by either (1) directly incorporating labels as input or (2) involving labels in an auxiliary classification loss. In this paper, we show that the former directly aligns the class-conditioned fake-and-real data distributions $P(\text{image}|\text{class})$ ({\em data matching}), while the latter aligns data-conditioned class distributions $P(\text{class}|\text{image})$ ({\em label matching}). Although class separability does not directly translate to sample quality and becomes a burden if classification itself is intrinsically difficult, the discriminator cannot provide useful guidance for the generator if features of distinct classes are mapped to the same point and thus become inseparable. Motivated by this intuition, we propose a Dual Projection GAN (P2GAN) model that learns to balance between {\em data matching} and {\em label matching}. We then propose an improved cGAN model with Auxiliary Classification that directly aligns the fake and real conditionals $P(\text{class}|\text{image})$ by minimizing their $f$-divergence. Experiments on a synthetic Mixture of Gaussian (MoG) dataset and a variety of real-world datasets including CIFAR100, ImageNet, and VGGFace2 demonstrate the efficacy of our proposed models.
We propose a communication-efficient collaborative inference framework in the domain of edge inference, focusing on the efficient use of vision transformer (ViT) models. The partitioning strategy of conventional collaborative inference fails to reduce communication cost because of the inherent architecture of ViTs maintaining consistent layer dimensions across the entire transformer encoder. Therefore, instead of employing the partitioning strategy, our framework utilizes a lightweight ViT model on the edge device, with the server deploying a complicated ViT model. To enhance communication efficiency and achieve the classification accuracy of the server model, we propose two strategies: 1) attention-aware patch selection and 2) entropy-aware image transmission. Attention-aware patch selection leverages the attention scores generated by the edge device's transformer encoder to identify and select the image patches critical for classification. This strategy enables the edge device to transmit only the essential patches to the server, significantly improving communication efficiency. Entropy-aware image transmission uses min-entropy as a metric to accurately determine whether to depend on the lightweight model on the edge device or to request the inference from the server model. In our framework, the lightweight ViT model on the edge device acts as a semantic encoder, efficiently identifying and selecting the crucial image information required for the classification task. Our experiments demonstrate that the proposed collaborative inference framework can reduce communication overhead by 68% with only a minimal loss in accuracy compared to the server model on the ImageNet dataset.
In this paper, we propose the differentiable channel sparsity search (DCSS) for convolutional neural networks. Unlike traditional channel pruning algorithms which require users to manually set prune ratios for each convolutional layer, DCSS automatically searches the optimal combination of sparsities. Inspired by the differentiable architecture search (DARTS), we draw lessons from the continuous relaxation and leverage the gradient information to balance the computational cost and metrics. Since directly applying the scheme of DARTS causes shape mismatching and excessive memory consumption, we introduce a novel technique called weight sharing within filters. This technique elegantly eliminates the problem of shape mismatching with negligible additional resources. We conduct comprehensive experiments on not only image classification but also find-grained tasks including semantic segmentation and image super resolution to verify the effectiveness of DCSS. Compared with previous network pruning approaches, DCSS achieves state-of-the-art results for image classification. Experimental results of semantic segmentation and image super resolution indicate that task-specific search achieves better performance than transferring slim models, demonstrating the wide applicability and high efficiency of DCSS.
Convolutional Neural Networks are particularly suited for image analysis tasks, such as Image Classification, Object Recognition or Image Segmentation. Like all Artificial Neural Networks, however, they are "black box" models, and suffer from poor explainability. This work is concerned with the specific downstream task of Emotion Recognition from images, and proposes a framework that combines CAM-based techniques with Object Detection on a corpus level to better understand on which image cues a particular model, in our case EmoNet, relies to assign a specific emotion to an image. We demonstrate that the model mostly focuses on human characteristics, but also explore the pronounced effect of specific image modifications.
The project aims to develop differentially private deep learning models for image classification on CIFAR-10 datasets \cite{cifar10} and analyze the impact of various privacy parameters on model accuracy. We have implemented five different deep learning models, namely ConvNet, ResNet18, EfficientNet, ViT, and DenseNet121 and three supervised classifiers namely K-Nearest Neighbors, Naive Bayes Classifier and Support Vector Machine. We evaluated the performance of these models under varying settings. Our best performing model to date is EfficientNet with test accuracy of $59.63\%$ with the following parameters (Adam optimizer, batch size 256, epoch size 100, epsilon value 5.0, learning rate $1e-3$, clipping threshold 1.0, and noise multiplier 0.912).
TextCNN, the convolutional neural network for text, is a useful deep learning algorithm for sentence classification tasks such as sentiment analysis and question classification. However, neural networks have long been known as black boxes because interpreting them is a challenging task. Researchers have developed several tools to understand a CNN for image classification by deep visualization, but research about deep TextCNNs is still insufficient. In this paper, we are trying to understand what a TextCNN learns on two classical NLP datasets. Our work focuses on functions of different convolutional kernels and correlations between convolutional kernels.
The goal of lossy data compression is to reduce the storage cost of a data set $X$ while retaining as much information as possible about something ($Y$) that you care about. For example, what aspects of an image $X$ contain the most information about whether it depicts a cat? Mathematically, this corresponds to finding a mapping $X\to Z\equiv f(X)$ that maximizes the mutual information $I(Z,Y)$ while the entropy $H(Z)$ is kept below some fixed threshold. We present a method for mapping out the Pareto frontier for classification tasks, reflecting the tradeoff between retained entropy and class information. We first show how a random variable $X$ (an image, say) drawn from a class $Y\in\{1,...,n\}$ can be distilled into a vector $W=f(X)\in \mathbb{R}^{n-1}$ losslessly, so that $I(W,Y)=I(X,Y)$; for example, for a binary classification task of cats and dogs, each image $X$ is mapped into a single real number $W$ retaining all information that helps distinguish cats from dogs. For the $n=2$ case of binary classification, we then show how $W$ can be further compressed into a discrete variable $Z=g_\beta(W)\in\{1,...,m_\beta\}$ by binning $W$ into $m_\beta$ bins, in such a way that varying the parameter $\beta$ sweeps out the full Pareto frontier, solving a generalization of the Discrete Information Bottleneck (DIB) problem. We argue that the most interesting points on this frontier are "corners" maximizing $I(Z,Y)$ for a fixed number of bins $m=2,3...$ which can be conveniently be found without multiobjective optimization. We apply this method to the CIFAR-10, MNIST and Fashion-MNIST datasets, illustrating how it can be interpreted as an information-theoretically optimal image clustering algorithm.
Spike-based neuromorphic hardware promises to reduce the energy consumption of image classification and other deep learning applications, particularly on mobile phones or other edge devices. However, direct training of deep spiking neural networks is difficult, and previous methods for converting trained artificial neural networks to spiking neurons were inefficient because the neurons had to emit too many spikes. We show that a substantially more efficient conversion arises when one optimizes the spiking neuron model for that purpose, so that it not only matters for information transmission how many spikes a neuron emits, but also when it emits those spikes. This advances the accuracy that can be achieved for image classification with spiking neurons, and the resulting networks need on average just two spikes per neuron for classifying an image. In addition, our new conversion method improves latency and throughput of the resulting spiking networks.
Pixel binning is a technique, widely used in optical image acquisition and spectroscopy, in which adjacent detector elements of an image sensor are combined into larger pixels. This reduces the amount of data to be processed as well as the impact of noise, but comes at the cost of a loss of information. Here, we push the concept of binning to its limit by combining a large fraction of the sensor elements into a single superpixel that extends over the whole face of the chip. For a given pattern recognition task, its optimal shape is determined from training data using a machine learning algorithm. We demonstrate the classification of optically projected images from the MNIST dataset on a nanosecond timescale, with enhanced sensitivity and without loss of classification accuracy. Our concept is not limited to imaging alone but can also be applied in optical spectroscopy or other sensing applications.
Language-vision models like CLIP have made significant strides in vision tasks, such as zero-shot image classification (ZSIC). However, generating specific and expressive visual descriptions remains challenging; descriptions produced by current methods are often ambiguous and lacking in granularity. To tackle these issues, we propose V-GLOSS: Visual Glosses, a novel method built upon two key ideas. The first is Semantic Prompting, which conditions a language model on structured semantic knowledge. The second is a new contrastive algorithm that elicits fine-grained distinctions between similar concepts. With both ideas, we demonstrate that V-GLOSS improves visual descriptions and achieves strong results in the zero-shot setting on general and fine-grained image-classification datasets, including ImageNet, STL-10, FGVC Aircraft, and Flowers 102. Moreover, these descriptive capabilities contribute to enhancing image-generation performance. Finally, we introduce a quality-tested silver dataset with descriptions generated with V-GLOSS for all ImageNet classes.
Few-shot semantic segmentation aims to recognize novel classes with only very few labelled data. This challenging task requires mining of the relevant relationships between the query image and the support images. Previous works have typically regarded it as a pixel-wise classification problem. Therefore, various models have been designed to explore the correlation of pixels between the query image and the support images. However, they focus only on pixel-wise correspondence and ignore the overall correlation of objects. In this paper, we introduce a mask-based classification method for addressing this problem. The mask aggregation network (MANet), which is a simple mask classification model, is proposed to simultaneously generate a fixed number of masks and their probabilities of being targets. Then, the final segmentation result is obtained by aggregating all the masks according to their locations. Experiments on both the PASCAL-5^i and COCO-20^i datasets show that our method performs comparably to the state-of-the-art pixel-based methods. This competitive performance demonstrates the potential of mask classification as an alternative baseline method in few-shot semantic segmentation. Our source code will be made available at https://github.com/TinyAway/MANet.
Medical data often exhibits distribution shifts, which cause test-time performance degradation for deep learning models trained using standard supervised learning pipelines. This challenge is addressed in the field of Domain Generalization (DG) with the sub-field of Single Domain Generalization (SDG) being specifically interesting due to the privacy- or logistics-related issues often associated with medical data. Existing disentanglement-based SDG methods heavily rely on structural information embedded in segmentation masks, however classification labels do not provide such dense information. This work introduces a novel SDG method aimed at medical image classification that leverages channel-wise contrastive disentanglement. It is further enhanced with reconstruction-based style regularization to ensure extraction of distinct style and structure feature representations. We evaluate our method on the complex task of multicenter histopathology image classification, comparing it against state-of-the-art (SOTA) SDG baselines. Results demonstrate that our method surpasses the SOTA by a margin of 1% in average accuracy while also showing more stable performance. This study highlights the importance and challenges of exploring SDG frameworks in the context of the classification task. The code is publicly available at https://github.com/BioMedIA-MBZUAI/ConDiSR
We explore the problem of classification within a medical image data-set based on a feature vector extracted from the deepest layer of pre-trained Convolution Neural Networks. We have used feature vectors from several pre-trained structures, including networks with/without transfer learning to evaluate the performance of pre-trained deep features versus CNNs which have been trained by that specific dataset as well as the impact of transfer learning with a small number of samples. All experiments are done on Kimia Path24 dataset which consists of 27,055 histopathology training patches in 24 tissue texture classes along with 1,325 test patches for evaluation. The result shows that pre-trained networks are quite competitive against training from scratch. As well, fine-tuning does not seem to add any tangible improvement for VGG16 to justify additional training while we observed considerable improvement in retrieval and classification accuracy when we fine-tuned the Inception structure.
Recently, deep learning models have achieved great success in computer vision applications, relying on large-scale class-balanced datasets. However, imbalanced class distributions still limit the wide applicability of these models due to degradation in performance. To solve this problem, in this paper, we concentrate on the study of cross entropy which mostly ignores output scores on incorrect classes. This work discovers that neutralizing predicted probabilities on incorrect classes improves the prediction accuracy for imbalanced image classification. This paper proposes a simple but effective loss named complement cross entropy based on this finding. The proposed loss makes the ground truth class overwhelm the other classes in terms of softmax probability, by neutralizing probabilities of incorrect classes, without additional training procedures. Along with it, this loss facilitates the models to learn key information especially from samples on minority classes. It ensures more accurate and robust classification results on imbalanced distributions. Extensive experiments on imbalanced datasets demonstrate the effectiveness of the proposed method.
Despite the unresolved 'dying ReLU problem', the classical ReLU activation function (AF) has been extensively applied in Deep Neural Networks (DNN), in particular Convolutional Neural Networks (CNN), for image classification. The common gradient issues of ReLU pose challenges in applications on academy and industry sectors. Recent approaches for improvements are in a similar direction by just proposing variations of the AF, such as Leaky ReLU (LReLU), while maintaining the solution within the same unresolved gradient problems. In this paper, the Absolute Leaky ReLU (ALReLU) AF, a variation of LReLU, is proposed, as an alternative method to resolve the common 'dying ReLU problem' on NN-based algorithms for supervised learning. The experimental results demonstrate that by using the absolute values of LReLU's small negative gradient, has a significant improvement in comparison with LReLU and ReLU, on image classification of diseases such as COVID-19, text and tabular data classification tasks on five different datasets.
Deep Convolutional Neural Networks (CNNs) have long been the architecture of choice for computer vision tasks. Recently, Transformer-based architectures like Vision Transformer (ViT) have matched or even surpassed ResNets for image classification. However, details of the Transformer architecture -- such as the use of non-overlapping patches -- lead one to wonder whether these networks are as robust. In this paper, we perform an extensive study of a variety of different measures of robustness of ViT models and compare the findings to ResNet baselines. We investigate robustness to input perturbations as well as robustness to model perturbations. We find that when pre-trained with a sufficient amount of data, ViT models are at least as robust as the ResNet counterparts on a broad range of perturbations. We also find that Transformers are robust to the removal of almost any single layer, and that while activations from later layers are highly correlated with each other, they nevertheless play an important role in classification.