Part-prototype models are explainable-by-design image classifiers, and a promising alternative to black box AI. This paper explores the applicability and potential of interpretable machine learning, in particular PIP-Net, for automated diagnosis support on real-world medical imaging data. PIP-Net learns human-understandable prototypical image parts and we evaluate its accuracy and interpretability for fracture detection and skin cancer diagnosis. We find that PIP-Net's decision making process is in line with medical classification standards, while only provided with image-level class labels. Because of PIP-Net's unsupervised pretraining of prototypes, data quality problems such as undesired text in an X-ray or labelling errors can be easily identified. Additionally, we are the first to show that humans can manually correct the reasoning of PIP-Net by directly disabling undesired prototypes. We conclude that part-prototype models are promising for medical applications due to their interpretability and potential for advanced model debugging.
Recent advancements in deep learning have shown transformative potential in medical imaging, yet concerns about fairness persist due to performance disparities across demographic subgroups. Existing methods aim to address these biases by mitigating sensitive attributes in image data; however, these attributes often carry clinically relevant information, and their removal can compromise model performance-a highly undesirable outcome. To address this challenge, we propose Fair Re-fusion After Disentanglement (FairREAD), a novel, simple, and efficient framework that mitigates unfairness by re-integrating sensitive demographic attributes into fair image representations. FairREAD employs orthogonality constraints and adversarial training to disentangle demographic information while using a controlled re-fusion mechanism to preserve clinically relevant details. Additionally, subgroup-specific threshold adjustments ensure equitable performance across demographic groups. Comprehensive evaluations on a large-scale clinical X-ray dataset demonstrate that FairREAD significantly reduces unfairness metrics while maintaining diagnostic accuracy, establishing a new benchmark for fairness and performance in medical image classification.
In recent years, Convolutional Neural Networks (CNNs) have shown superior capability in visual learning tasks. While accuracy-wise CNNs provide unprecedented performance, they are also known to be computationally intensive and energy demanding for modern computer systems. In this paper, we propose Virtual Pooling (ViP), a model-level approach to improve speed and energy consumption of CNN-based image classification and object detection tasks, with a provable error bound. We show the efficacy of ViP through experiments on four CNN models, three representative datasets, both desktop and mobile platforms, and two visual learning tasks, i.e., image classification and object detection. For example, ViP delivers 2.1x speedup with less than 1.5% accuracy degradation in ImageNet classification on VGG-16, and 1.8x speedup with 0.025 mAP degradation in PASCAL VOC object detection with Faster-RCNN. ViP also reduces mobile GPU and CPU energy consumption by up to 55% and 70%, respectively. As a complementary method to existing acceleration approaches, ViP achieves 1.9x speedup on ThiNet leading to a combined speedup of 5.23x on VGG-16. Furthermore, ViP provides a knob for machine learning practitioners to generate a set of CNN models with varying trade-offs between system speed/energy consumption and accuracy to better accommodate the requirements of their tasks. Code is available at https://github.com/cmu-enyac/VirtualPooling.
For image segmentation, the current standard is to perform pixel-level optimization and inference in Euclidean output embedding spaces through linear hyperplanes. In this work, we show that hyperbolic manifolds provide a valuable alternative for image segmentation and propose a tractable formulation of hierarchical pixel-level classification in hyperbolic space. Hyperbolic Image Segmentation opens up new possibilities and practical benefits for segmentation, such as uncertainty estimation and boundary information for free, zero-label generalization, and increased performance in low-dimensional output embeddings.
Unsupervised methods have proven effective for discriminative tasks in a single-modality scenario. In this paper, we present a multimodal framework for learning sparse representations that can capture semantic correlation between modalities. The framework can model relationships at a higher level by forcing the shared sparse representation. In particular, we propose the use of joint dictionary learning technique for sparse coding and formulate the joint representation for concision, cross-modal representations (in case of a missing modality), and union of the cross-modal representations. Given the accelerated growth of multimodal data posted on the Web such as YouTube, Wikipedia, and Twitter, learning good multimodal features is becoming increasingly important. We show that the shared representations enabled by our framework substantially improve the classification performance under both unimodal and multimodal settings. We further show how deep architectures built on the proposed framework are effective for the case of highly nonlinear correlations between modalities. The effectiveness of our approach is demonstrated experimentally in image denoising, multimedia event detection and retrieval on the TRECVID dataset (audio-video), category classification on the Wikipedia dataset (image-text), and sentiment classification on PhotoTweet (image-text).
With the development of convolutional neural networks (CNNs) in recent years, the network structure has become more and more complex and varied, and has achieved very good results in pattern recognition, image classification, object detection and tracking. For CNNs used for image classification, in addition to the network structure, more and more research is now focusing on the improvement of the loss function, so as to enlarge the inter-class feature differences, and reduce the intra-class feature variations as soon as possible. Besides the traditional Softmax, typical loss functions include L-Softmax, AM-Softmax, ArcFace, and Center loss, etc. Based on the concept of predefined evenly-distributed class centroids (PEDCC) in CSAE network, this paper proposes a PEDCC-based loss function called PEDCC-Loss, which can make the inter-class distance maximal and intra-class distance small enough in hidden feature space. Multiple experiments on image classification and face recognition have proved that our method achieve the best recognition accuracy, and network training is stable and easy to converge. Code is available in https://github.com/ZLeopard/PEDCC-Loss
In order to extract knowledge from the large data collected by edge devices, traditional cloud based approach that requires data upload may not be feasible due to communication bandwidth limitation as well as privacy and security concerns of end users. To address these challenges, a novel privacy preserving edge computing framework is proposed in this paper for image classification. Specifically, autoencoder will be trained unsupervised at each edge device individually, then the obtained latent vectors will be transmitted to the edge server for the training of a classifier. This framework would reduce the communications overhead and protect the data of the end users. Comparing to federated learning, the training of the classifier in the proposed framework does not subject to the constraints of the edge devices, and the autoencoder can be trained independently at each edge device without any server involvement. Furthermore, the privacy of the end users' data is protected by transmitting latent vectors without additional cost of encryption. Experimental results provide insights on the image classification performance vs. various design parameters such as the data compression ratio of the autoencoder and the model complexity.
Some recent methods address few-shot image classification by extracting semantic information from class names and devising mechanisms for aligning vision and semantics to integrate information from both modalities. However, class names provide only limited information, which is insufficient to capture the visual details in images. As a result, such vision-semantics alignment is inherently biased, leading to suboptimal integration outcomes. In this paper, we avoid such biased vision-semantics alignment by introducing CLIP, a natural bridge between vision and semantics, and enforcing unbiased vision-vision alignment as a proxy task. Specifically, we align features encoded from the few-shot encoder and CLIP's vision encoder on the same image. This alignment is accomplished through a linear projection layer, with a training objective formulated using optimal transport-based assignment prediction. Thanks to...
Medical Visual Question Answering (VQA) systems play a supporting role to understand clinic-relevant information carried by medical images. The questions to a medical image include two categories: close-end (such as Yes/No question) and open-end. To obtain answers, the majority of the existing medical VQA methods relies on classification approaches, while a few works attempt to use generation approaches or a mixture of the two. The classification approaches are relatively simple but perform poorly on long open-end questions. To bridge this gap, in this paper, we propose a new Transformer based framework for medical VQA (named as Q2ATransformer), which integrates the advantages of both the classification and the generation approaches and provides a unified treatment for the close-end and open-end questions. Specifically, we introduce an additional Transformer decoder with a set of learnable candidate answer embeddings to query the existence of each answer class to a given image-question pair. Through the Transformer attention, the candidate answer embeddings interact with the fused features of the image-question pair to make the decision. In this way, despite being a classification-based approach, our method provides a mechanism to interact with the answer information for prediction like the generation-based approaches. On the other hand, by classification, we mitigate the task difficulty by reducing the search space of answers. Our method achieves new state-of-the-art performance on two medical VQA benchmarks. Especially, for the open-end questions, we achieve 79.19% on VQA-RAD and 54.85% on PathVQA, with 16.09% and 41.45% absolute improvements, respectively.
On basis of functional magnetic resonance imaging (fMRI), researchers are devoted to designing visual encoding models to predict the neuron activity of human in response to presented image stimuli and analyze inner mechanism of human visual cortices. Deep network structure composed of hierarchical processing layers forms deep network models by learning features of data on specific task through big dataset. Deep network models have powerful and hierarchical representation of data, and have brought about breakthroughs for visual encoding, while revealing hierarchical structural similarity with the manner of information processing in human visual cortices. However, previous studies almost used image features of those deep network models pre-trained on classification task to construct visual encoding models. Except for deep network structure, the task or corresponding big dataset is also important for deep network models, but neglected by previous studies. Because image classification is a relatively fundamental task, it is difficult to guide deep network models to master high-level semantic representations of data, which causes into that encoding performance for high-level visual cortices is limited. In this study, we introduced one higher-level vision task: image caption (IC) task and proposed the visual encoding model based on IC features (ICFVEM) to encode voxels of high-level visual cortices. Experiment demonstrated that ICFVEM obtained better encoding performance than previous deep network models pre-trained on classification task. In addition, the interpretation of voxels was realized to explore the detailed characteristics of voxels based on the visualization of semantic words, and comparative analysis implied that high-level visual cortices behaved the correlative representation of image content.
High-quality data is necessary for modern machine learning. However, the acquisition of such data is difficult due to noisy and ambiguous annotations of humans. The aggregation of such annotations to determine the label of an image leads to a lower data quality. We propose a data-centric image classification benchmark with ten real-world datasets and multiple annotations per image to allow researchers to investigate and quantify the impact of such data quality issues. With the benchmark we can study the impact of annotation costs and (semi-)supervised methods on the data quality for image classification by applying a novel methodology to a range of different algorithms and diverse datasets. Our benchmark uses a two-phase approach via a data label improvement method in the first phase and a fixed evaluation model in the second phase. Thereby, we give a measure for the relation between the input labeling effort and the performance of (semi-)supervised algorithms to enable a deeper insight into how labels should be created for effective model training. Across thousands of experiments, we show that one annotation is not enough and that the inclusion of multiple annotations allows for a better approximation of the real underlying class distribution. We identify that hard labels can not capture the ambiguity of the data and this might lead to the common issue of overconfident models. Based on the presented datasets, benchmarked methods, and analysis, we create multiple research opportunities for the future directed at the improvement of label noise estimation approaches, data annotation schemes, realistic (semi-)supervised learning, or more reliable image collection.
This paper have two parts. In the first part we discuss word embeddings. We discuss the need for them, some of the methods to create them, and some of their interesting properties. We also compare them to image embeddings and see how word embedding and image embedding can be combined to perform different tasks. In the second part we implement a convolutional neural network trained on top of pre-trained word vectors. The network is used for several sentence-level classification tasks, and achieves state-of-art (or comparable) results, demonstrating the great power of pre-trainted word embeddings over random ones.
Modern AI image classifiers have made impressive advances in recent years, but their performance often appears strange or violates expectations of users. This suggests humans engage in cognitive anthropomorphism: expecting AI to have the same nature as human intelligence. This mismatch presents an obstacle to appropriate human-AI interaction. To delineate this mismatch, I examine known properties of human classification, in comparison to image classifier systems. Based on this examination, I offer three strategies for system design that can address the mismatch between human and AI classification: explainable AI, novel methods for training users, and new algorithms that match human cognition.
CNNs exhibit many behaviors different from humans, one of which is the capability of employing high-frequency components. This paper discusses the frequency bias phenomenon in image classification tasks: the high-frequency components are actually much less exploited than the low- and mid-frequency components. We first investigate the frequency bias phenomenon by presenting two observations on feature discrimination and learning priority. Furthermore, we hypothesize that (i) the spectral density, (ii) class consistency directly affect the frequency bias. Specifically, our investigations verify that the spectral density of datasets mainly affects the learning priority, while the class consistency mainly affects the feature discrimination.
Environmental Microorganism Data Set Fifth Version (EMDS-5) is a microscopic image dataset including original Environmental Microorganism (EM) images and two sets of Ground Truth (GT) images. The GT image sets include a single-object GT image set and a multi-object GT image set. The EMDS-5 dataset has 21 types of EMs, each of which contains 20 original EM images, 20 single-object GT images and 20 multi-object GT images. EMDS-5 can realize to evaluate image preprocessing, image segmentation, feature extraction, image classification and image retrieval functions. In order to prove the effectiveness of EMDS-5, for each function, we select the most representative algorithms and price indicators for testing and evaluation. The image preprocessing functions contain two parts: image denoising and image edge detection. Image denoising uses nine kinds of filters to denoise 13 kinds of noises, respectively. In the aspect of edge detection, six edge detection operators are used to detect the edges of the images, and two evaluation indicators, peak-signal to noise ratio and mean structural similarity, are used for evaluation. Image segmentation includes single-object image segmentation and multi-object image segmentation. Six methods are used for single-object image segmentation, while k-means and U-net are used for multi-object segmentation.We extract nine features from the images in EMDS-5 and use the Support Vector Machine classifier for testing. In terms of image classification, we select the VGG16 feature to test different classifiers. We test two types of retrieval approaches: texture feature retrieval and deep learning feature retrieval. We select the last layer of features of these two deep learning networks as feature vectors. We use mean average precision as the evaluation index for retrieval.
Artificial neural networks (ANNs) have gained a well-deserved popularity among machine learning tools upon their recent successful applications in image- and sound processing and classification problems. ANNs have also been applied for predicting the family or function of a protein, knowing its residue sequence. Here we present two new ANNs with multi-label classification ability, showing impressive accuracy when classifying protein sequences into 698 UniProt families (AUC=99.99%) and 983 Gene Ontology classes (AUC=99.45%).
In this paper, we have proposed a new glaucoma classification approach that employs a wavelet neural network (WNN) on optimally enhanced retinal images features. To avoid tedious and error prone manual analysis of retinal images by ophthalmologists, computer aided diagnosis (CAD) substantially aids in robust diagnosis. Our objective is to introduce a CAD system with a fresh approach. Retinal image quality improvement is attempted in two phases. The retinal image preprocessing phase improves the brightness and contrast of the image through quantile based histogram modification. It is followed by the image enhancement phase, which involves multi scale morphological operations using image specific dynamic structuring elements for the retinal structure enrichment. Graph based retinal image features in terms of Local Graph Structures (LGS) and Graph Shortest Path (GSP) statistics are extracted from various directions along with the statistical features from the enhanced retinal dataset. WNN is employed to classify glaucoma retinal images with a suitable wavelet activation function. The performance of the WNN classifier is compared with multilayer perceptron neural networks with various datasets. The results show our approach is superior to the existing approaches.
Images and structured tables are essential parts of real-world databases. Though tabular-image representation learning is promising to create new insights, it remains a challenging task, as tabular data is typically heterogeneous and incomplete, presenting significant modality disparities with images. Earlier works have mainly focused on simple modality fusion strategies in complete data scenarios, without considering the missing data issue, and thus are limited in practice. In this paper, we propose TIP, a novel tabular-image pre-training framework for learning multimodal representations robust to incomplete tabular data. Specifically, TIP investigates a novel self-supervised learning (SSL) strategy, including a masked tabular reconstruction task for tackling data missingness, and image-tabular matching and contrastive learning objectives to capture multimodal information. Moreover, TIP proposes a versatile tabular encoder tailored for incomplete, heterogeneous tabular data and a multimodal interaction module for inter-modality representation learning. Experiments are performed on downstream multimodal classification tasks using both natural and medical image datasets. The results show that TIP outperforms state-of-the-art supervised/SSL image/multimodal algorithms in both complete and incomplete data scenarios. Our code is available at https://github.com/siyi-wind/TIP.
We propose a novel iterative method to adapt a a graph to d-dimensional image data. The method drives the nodes of the graph towards image features. The adaptation process naturally lends itself to a measure of feature saliency which can then be used to retain meaningful nodes and edges in the graph. From the adapted graph, we also propose the computation of a dual graph, which inherits the saliency measure from the adapted graph, and whose edges run along image features, hence producing an oversegmenting graph. The proposed method is computationally efficient and fully parallelisable. We propose two distance measures to find image saliency along graph edges, and evaluate the performance on synthetic images and on natural images from publicly available databases. In both cases, the most salient nodes of the graph achieve average boundary recall over 90%. We also apply our method to image classification on the MNIST hand-written digit dataset, using a recently proposed Deep Geometric Learning architecture, and achieving state-of-the-art classification accuracy, for a graph-based method, of 97.86%.
Recent advances in computer vision using deep learning with RGB imagery (e.g., object recognition and detection) have been made possible thanks to the development of large annotated RGB image datasets. In contrast, multispectral image (MSI) and hyperspectral image (HSI) datasets contain far fewer labeled images, in part due to the wide variety of sensors used. These annotations are especially limited for semantic segmentation, or pixel-wise classification, of remote sensing imagery because it is labor intensive to generate image annotations. Low-shot learning algorithms can make effective inferences despite smaller amounts of annotated data. In this paper, we study low-shot learning using self-taught feature learning for semantic segmentation. We introduce 1) an improved self-taught feature learning framework for HSI and MSI data and 2) a semi-supervised classification algorithm. When these are combined, they achieve state-of-the-art performance on remote sensing datasets that have little annotated training data available. These low-shot learning frameworks will reduce the manual image annotation burden and improve semantic segmentation performance for remote sensing imagery.