A new method compares image classifiers using adaptive sampling of natural images.
problem Evaluation of image classifiers on small, fixed test sets may not generalize to real-world images.
method Adaptive sampling from a large corpus of unlabeled images to maximize classifier discrepancies measured by WordNet hierarchy.
result Human labeling of model-dependent image sets reveals relative classifier performance.
Image classifiers are sensitive to small changes, affecting most images in a class.
problem Sensitivity of image classifiers to small perturbations.
method Demonstrated sensitivity for any classifier over images, showing that for most classes, a tiny perturbation can change the classification of a majority of images.
result Image classifiers are sensitive to small perturbations, affecting most images in a class.
Traditional classifiers can generate high-quality images comparable to generative models.
problem Separation between classifiers and generators in neural networks.
method Optimizing input gradients to produce images, using mask-based stochastic reconstruction, progressive-resolution technique, and distance metric loss.
result Traditional classifiers can generate high-fidelity images of 256imes256 resolution on ImageNet. StylEx trains a GAN to explain classifier decisions in StyleSpace.
problem Creating meaningful image-specific explanations for classifier decisions.
method Training a StyleGAN to learn a classifier-specific StyleSpace, incorporating the classifier model.
result StylEx finds attributes that align with semantic ones and generates human-interpretable explanations.
Understanding if classifiers generalize to out-of-sample datasets is a central problem in machine learning. Microscopy images provide a standardized way to measure the generalization capacity of image classifiers, as we can image the same classes of objects under increasingly divergent, but controlled factors of variat…
HET-XL improves heteroscedastic classifiers for large-scale image classification.
problem Scaling heteroscedastic classifiers to handle large numbers of classes and tuning the temperature hyperparameter.
method HET-XL, a heteroscedastic classifier with independent parameter count from the number of classes, learns the temperature hyperparameter directly from training data.
result HET-XL requires 14X fewer additional parameters and performs better than baseline heteroscedastic classifiers on large image classification datasets.
New attack tricks certifiably robust models into mislabeling images.
problem Defeating certified defenses against adversarial examples.
method Spoofed robustness certificates and large perturbations.
result Certifiably robust models can be fooled by large perturbations.
New image classifier uses hierarchical max-pooling with local pooling.
problem Improving image classification accuracy with variable spatial relationships.
method Introduces a hierarchical max-pooling model with additional local pooling for convolutional neural networks.
result Demonstrates improved performance in estimating image features.
New method generates realistic adversarial images with smaller perturbations.
problem Generating realistic adversarial images with small perturbations.
method Imposes spatial distortions (scaling, rotation, shear, translation) in addition to perturbations.
result Produces visually more realistic attacks that deceive classifiers without affecting human predictions.
Paper evaluates using app images for classification, improving accuracy.
problem Improving app classification accuracy when text descriptions are missing or inadequate.
method Used OCR, pic2vec, captionbot.ai, and object detection to convert images into text or vectors for classification.
result Improved classification accuracy of 96% for some app categories when images are added.
Sepsis is a life-threatening disease and one of the major causes of death in hospitals. Imaging of microcirculatory dysfunction is a promising approach for automated diagnosis of sepsis. We report a machine learning classifier capable of distinguishing non-septic and septic images from dark field microcirculation video…
Convolutional neural networks handle rotated image symmetries without dimensionality issues.
problem Binary image classification with rotational symmetry.
method Least squares plug-in classifiers based on convolutional neural networks under rotationally symmetric assumptions.
result Convolutional neural networks can circumvent the curse of dimensionality in binary image classification with rotational symmetry.
DeDUCE efficiently generates realistic counterfactuals for image classifiers.
problem Generating accurate counterfactual explanations for large image classifiers.
method Developed a new algorithm using spectral normalization to efficiently generate counterfactuals.
result Our algorithm consistently produces counterfactuals closer to the original inputs with comparable realism.
Study assesses deep neural networks' robustness in mammogram images.
problem Understanding deep neural networks' robustness in mammogram images for breast cancer screening.
method Measuring sensitivity to four perturbations and analyzing low-pass filtering effects.
result Mammogram image classifiers are sensitive to perturbations similar to natural images, but low-pass filtering degrades clinically meaningful features.
Study improves image classifier robustness to random p-norm corruptions.
problem Improving robustness of image classifiers to real-world imperceptible corruptions.
method Training and testing with random p-norm corruptions, evaluating robustness against different p-norms.
result Training with a combination of p-norm corruptions significantly improves robustness.
Denoised smoothing defends pretrained classifiers against adversarial attacks.
problem Adversarial attacks on pretrained classifiers.
method Prepending a denoiser to any off-the-shelf classifier using randomized smoothing.
result Guaranteed ℓp-robustness to adversarial examples without modifying the pretrained classifier. A novel approach stores encoded images as centroids and covariance matrices to improve classification accuracy with less memory.
problem Catastrophic forgetting and memory limitations in continual learning.
method Trains autoencoders with Neural Style Transfer to encode images, replay encoded episodes to avoid forgetting, and use centroids and covariance matrices for pseudo-images when memory is full.
result Increases classification accuracy by 13-17% over state-of-the-art methods on benchmark datasets, while requiring 78% less storage space.
Today's state-of-the-art image classifiers fail to correctly classify carefully manipulated adversarial images. In this work, we develop a new, localized adversarial attack that generates adversarial examples by imperceptibly altering the backgrounds of normal images. We first use this attack to highlight the unnecessa…
EdgeFool generates adversarial images to mislead classifiers.
problem Misleading classifiers with adversarial images.
method Trains a fully convolutional neural network to generate perturbations that enhance image details and mislead classifiers.
result EdgeFool outperforms other adversarial methods on various classifiers and datasets.
Supervised learning has been very successful for automatic segmentation of images from a single scanner. However, several papers report deteriorated performances when using classifiers trained on images from one scanner to segment images from other scanners. We propose a transfer learning classifier that adapts to diff…
We address the problem of \emph{instance label stability} in multiple instance learning (MIL) classifiers. These classifiers are trained only on globally annotated images (bags), but often can provide fine-grained annotations for image pixels or patches (instances). This is interesting for computer aided diagnosis (CAD…
AdvFlow generates imperceptible adversarial images using normalizing flows.
problem Adversarial attacks on deep learning classifiers.
method Normalizing flows to model adversarial examples.
result Generated adversaries closely follow clean data distribution, making them harder to detect.
Diffusion models outperform GANs in image synthesis quality.
problem Improving image synthesis quality using generative models.
method Developed and optimized diffusion models, and introduced classifier guidance for conditional synthesis.
result Achieved superior image synthesis quality with diffusion models, matching or exceeding GANs in FID scores.
With the advents of deep learning, improved image classification with complex discriminative models has been made possible. However, such deep models with increased complexity require a huge set of labeled samples to generalize the training. Such classification models can easily overfit when applied for medical images …
We propose an approach to distinguish between correct and incorrect image classifications. Our approach can detect misclassifications which either occur unintentionally ("natural errors"), or due to intentional adversarial attacks ("adversarial errors"), both in a single unified framework. Our appr…
Counting and classifying blood cells is an important diagnostic tool in medicine. Support Vector Machines are increasingly popular and efficient and could replace artificial neural network systems. Here a method to classify blood cells is proposed using SVM. A set of statistics on images are implemented in C++. The MPE…
Generative models improve image classifier explanations by realistically removing features.
problem Perturbation-based explanations often produce unrealistic counterfactual samples.
method Integrate generative inpainters into attribution methods to remove input features.
result Improved explanation methods in object localization, deletion, and saliency metrics.
Convolutional neural networks improve image classification accuracy.
problem Improving accuracy in image classification.
method Analyzing the convergence rate of misclassification risk for image classifiers.
result A rate of convergence independent of image dimension proves the effectiveness of CNNs.
A deep learning approach classifies medical images hierarchically.
problem Limitations of traditional supervised classifiers in medical image classification.
method Hierarchical Medical Image Classification (HMIC) using deep learning models.
result HMIC achieved better performance in classifying medical images hierarchically.
In this work we develop a fast saliency detection method that can be applied to any differentiable image classifier. We train a masking model to manipulate the scores of the classifier by masking salient parts of the input image. Our model generalises well to unseen images and requires a single forward pass to perform …
Contextual PDA improves explanation of image classifications for saturated models.
problem Difficulty in explaining decisions of saturated classifiers.
method Proposes Contextual PDA, a faster method for explaining image classifications.
result Contextual PDA outperforms PDA in explaining image classifications of state-of-the-art deep networks.
A simple method flags images as out-of-distribution based on their distance to nearest neighbors.
problem Detecting images not aligned with a trained model's in-distribution data.
method Flag images as OOD if their average distance to K nearest neighbors is large in the classifier's representation space.
result Simple methods can outperform more complex ones when considering learned representations.
Automatically classifying the tissues types of Region of Interest (ROI) in medical imaging has been an important application in Computer-Aided Diagnosis (CAD), such as classification of breast parenchymal tissue in the mammogram, classify lung disease patterns in High-Resolution Computed Tomography (HRCT) etc. Recently…
This paper addresses detection of a reverse engineering (RE) attack targeting a deep neural network (DNN) image classifier; by querying, RE's aim is to discover the classifier's decision rule. RE can enable test-time evasion attacks, which require knowledge of the classifier. Recently, we proposed a quite effective app…
New method provides formal uncertainty guarantees for image classifiers.
problem Uncertainty quantification for image classifiers without formal guarantees.
method Adapts conformal prediction to give stable, formal coverage guarantees.
result Method outperforms existing approaches in coverage and set size.
Proposes a method to quantify uncertainty in deterministic image classifiers.
problem Uncertainty in deterministic image classifiers.
method Introduces Wellington Posterior for inductive transfer from scenes.
result Validates Wellington Posterior using various methods.
Facial analysis models are increasingly used in applications that have serious impacts on people's lives, ranging from authentication to surveillance tracking. It is therefore critical to develop techniques that can reveal unintended biases in facial classifiers to help guide the ethical use of facial analysis technolo…
Unified framework for classification and image generation.
problem Limited supervision in classification and image generation.
method Triple Generative Adversarial Network (Triple-GAN) as a three-player minimax game.
result Unique equilibrium converges to data distribution, achieving excellent classification and generation results.
We study the robustness of image classifiers to temporal perturbations derived from videos. As part of this study, we construct two datasets, ImageNet-Vid-Robust and YTBB-Robust , containing a total 57,897 images grouped into 3,139 sets of perceptually similar images. Our datasets were derived from ImageNet-Vid and You…
Images seen during test time are often not from the same distribution as images used for learning. This problem, known as domain shift, occurs when training classifiers from object-centric internet image databases and trying to apply them directly to scene understanding tasks. The consequence is often severe performanc…
Develops new methods to create imperceptible image changes that fool classifiers.
problem Improving the robustness of image classifiers by creating subtle changes undetectable to humans.
method Two methods: Edge-Aware and Color-Aware, designed to reduce detectability of image perturbations.
result Demonstrated that the new methods effectively cause misclassification and are computationally efficient.
New framework classifies generated images into finer categories.
problem Difficulty in detecting new types of generated images.
method Scalable deep metric learning for multi-class classification.
result Improved detection performance on new types of generated data.
ECP method improves image classifier uncertainty sets.
problem Generating reliable uncertainty sets for deep classifiers.
method Evidential Conformal Prediction (ECP) based on EDL.
result ECP outperforms state-of-the-art methods in set size and adaptivity.
Deep learning models designed for visual classification tasks on natural images have become prevalent in medical image analysis. However, medical images differ from typical natural images in many ways, such as significantly higher resolutions and smaller regions of interest. Moreover, both the global structure and loca…
We study the problem of finding a universal (image-agnostic) perturbation to fool machine learning (ML) classifiers (e.g., neural nets, decision tress) in the hard-label black-box setting. Recent work in adversarial ML in the white-box setting (model parameters are known) has shown that many state-of-the-art image clas…
The artistic style of a painting is a subtle aesthetic judgment used by art historians for grouping and classifying artwork. The recently introduced `neural-style' algorithm substantially succeeds in merging the perceived artistic style of one image or set of images with the perceived content of another. In light of th…
This research discovers model architecture and training dataset characteristics through strategic input probing.
problem Discovering model architecture and training dataset characteristics in black box models.
method Structured input probes and model outputs are used to train a deep classifier for image and text classification.
result The approach successfully distinguishes between different image and text datasets and architectures.
The paper uses VGG-19 for plant species classification from leaf images.
problem Manual inspection of plant species by botanists is time-consuming.
method Transfer learning with VGG-19 for feature extraction and classification.
result The model achieves 99.70% accuracy in predicting plant species.