This paper presents a simple and effective approach to solving the multi-label classification problem. The proposed approach leverages Transformer decoders to query the existence of a class label. The use of Transformer is rooted in the need of extracting local discriminative features adaptively for different labels, which is a strongly desired property due to the existence of multiple objects in one image. The built-in cross-attention module in the Transformer decoder offers an effective way to use label embeddings as queries to probe and pool class-related features from a feature map computed by a vision backbone for subsequent binary classifications. Compared with prior works, the new framework is simple, using standard Transformers and vision backbones, and effective, consistently outperforming all previous works on five multi-label classification data sets, including MS-COCO, PASCAL VOC, NUS-WIDE, and Visual Genome. Particularly, we establish $91.3\%$ mAP on MS-COCO. We hope its compact structure, simple implementation, and superior performance serve as a strong baseline for multi-label classification tasks and future studies. The code will be available soon at https://github.com/SlongLiu/query2labels.
The application of machine learning techniques to the medical domain is especially challenging due to the required level of precision and the incurrence of huge risks of minute errors. Employing these techniques to a more complex subdomain of hematological diagnosis seems quite promising, with automatic identification of blood cell types, which can help in detection of hematologic disorders. In this paper, we benchmark 27 popular deep convolutional neural network architectures on the microscopic peripheral blood cell images dataset. The dataset is publicly available, with large number of normal peripheral blood cells acquired using the CellaVision DM96 analyzer and identified by expert pathologists into eight different cell types. We fine-tune the state-of-the-art image classification models pre-trained on the ImageNet dataset for blood cell classification. We exploit data augmentation techniques during training to avoid overfitting and achieve generalization. An ensemble of the top performing models obtains significant improvements over past published works, achieving the state-of-the-art results with a classification accuracy of 99.51%. Our work provides empirical baselines and benchmarks on standard deep-learning architectures for microscopic peripheral blood cell recognition task.
Negative flips are errors introduced in a classification system when a legacy model is updated. Existing methods to reduce the negative flip rate (NFR) either do so at the expense of overall accuracy by forcing a new model to imitate the old models, or use ensembles, which multiply inference cost prohibitively. We analyze the role of ensembles in reducing NFR and observe that they remove negative flips that are typically not close to the decision boundary, but often exhibit large deviations in the distance among their logits. Based on the observation, we present a method, called Ensemble Logit Difference Inhibition (ELODI), to train a classification system that achieves paragon performance in both error rate and NFR, at the inference cost of a single model. The method distills a homogeneous ensemble to a single student model which is used to update the classification system. ELODI also introduces a generalized distillation objective, Logit Difference Inhibition (LDI), which only penalizes the logit difference of a subset of classes with the highest logit values. On multiple image classification benchmarks, model updates with ELODI demonstrate superior accuracy retention and NFR reduction.
Advancements in deep learning techniques have given a boost to the performance of anomaly detection. However, real-world and safety-critical applications demand a level of transparency and reasoning beyond accuracy. The task of anomaly detection (AD) focuses on finding whether a given sample follows the learned distribution. Existing methods lack the ability to reason with clear explanations for their outcomes. Hence to overcome this challenge, we propose Transparent {A}nomaly Detection {C}oncept {E}xplanations (ACE). ACE is able to provide human interpretable explanations in the form of concepts along with anomaly prediction. To the best of our knowledge, this is the first paper that proposes interpretable by-design anomaly detection. In addition to promoting transparency in AD, it allows for effective human-model interaction. Our proposed model shows either higher or comparable results to black-box uninterpretable models. We validate the performance of ACE across three realistic datasets - bird classification on CUB-200-2011, challenging histopathology slide image classification on TIL-WSI-TCGA, and gender classification on CelebA. We further demonstrate that our concept learning paradigm can be seamlessly integrated with other classification-based AD methods.
Continual Learning (CL) is crucial for enabling networks to dynamically adapt as they learn new tasks sequentially, accommodating new data and classes without catastrophic forgetting. Diverging from conventional perspectives on CL, our paper introduces a new perspective wherein forgetting could actually benefit the sequential learning paradigm. Specifically, we present BiasPruner, a CL framework that intentionally forgets spurious correlations in the training data that could lead to shortcut learning. Utilizing a new bias score that measures the contribution of each unit in the network to learning spurious features, BiasPruner prunes those units with the highest bias scores to form a debiased subnetwork preserved for a given task. As BiasPruner learns a new task, it constructs a new debiased subnetwork, potentially incorporating units from previous subnetworks, which improves adaptation and performance on the new task. During inference, BiasPruner employs a simple task-agnostic approach to select the best debiased subnetwork for predictions. We conduct experiments on three medical datasets for skin lesion classification and chest X-Ray classification and demonstrate that BiasPruner consistently outperforms SOTA CL methods in terms of classification performance and fairness. Our code is available here.
Generative adversarial networks (GANs) have ushered in a revolution in image-to-image translation. The development and proliferation of GANs raises an interesting question: can we train a GAN to remove an object, if present, from an image while otherwise preserving the image? Specifically, can a GAN "virtually heal" anyone by turning his medical image, with an unknown health status (diseased or healthy), into a healthy one, so that diseased regions could be revealed by subtracting those two images? Such a task requires a GAN to identify a minimal subset of target pixels for domain translation, an ability that we call fixed-point translation, which no GAN is equipped with yet. Therefore, we propose a new GAN, called Fixed-Point GAN, trained by (1) supervising same-domain translation through a conditional identity loss, and (2) regularizing cross-domain translation through revised adversarial, domain classification, and cycle consistency loss. Based on fixed-point translation, we further derive a novel framework for disease detection and localization using only image-level annotation. Qualitative and quantitative evaluations demonstrate that the proposed method outperforms the state of the art in multi-domain image-to-image translation and that it surpasses predominant weakly-supervised localization methods in both disease detection and localization. Implementation is available at https://github.com/jlianglab/Fixed-Point-GAN.
Vision-and-Language Pre-training (VLP) improves model performance for downstream tasks that require image and text inputs. Current VLP approaches differ on (i) model architecture (especially image embedders), (ii) loss functions, and (iii) masking policies. Image embedders are either deep models like ResNet or linear projections that directly feed image-pixels into the transformer. Typically, in addition to the Masked Language Modeling (MLM) loss, alignment-based objectives are used for cross-modality interaction, and RoI feature regression and classification tasks for Masked Image-Region Modeling (MIRM). Both alignment and MIRM objectives mostly do not have ground truth. Alignment-based objectives require pairings of image and text and heuristic objective functions. MIRM relies on object detectors. Masking policies either do not take advantage of multi-modality or are strictly coupled with alignments generated by other models. In this paper, we present Masked Language and Image Modeling (MLIM) for VLP. MLIM uses two loss functions: Masked Language Modeling (MLM) loss and image reconstruction (RECON) loss. We propose Modality Aware Masking (MAM) to boost cross-modality interaction and take advantage of MLM and RECON losses that separately capture text and image reconstruction quality. Using MLM + RECON tasks coupled with MAM, we present a simplified VLP methodology and show that it has better downstream task performance on a proprietary e-commerce multi-modal dataset.
Convolutional Neural Networks (CNNs) have had great success in many machine vision as well as machine audition tasks. Many image recognition network architectures have consequently been adapted for audio processing tasks. However, despite some successes, the performance of many of these did not translate from the image to the audio domain. For example, very deep architectures such as ResNet and DenseNet, which significantly outperform VGG in image recognition, do not perform better in audio processing tasks such as Acoustic Scene Classification (ASC). In this paper, we investigate the reasons why such powerful architectures perform worse in ASC compared to simpler models (e.g., VGG). To this end, we analyse the receptive field (RF) of these CNNs and demonstrate the importance of the RF to the generalization capability of the models. Using our receptive field analysis, we adapt both ResNet and DenseNet, achieving state-of-the-art performance and eventually outperforming the VGG-based models. We introduce systematic ways of adapting the RF in CNNs, and present results on three data sets that show how changing the RF over the time and frequency dimensions affects a model's performance. Our experimental results show that very small or very large RFs can cause performance degradation, but deep models can be made to generalize well by carefully choosing an appropriate RF size within a certain range.
We propose a generalized Sparse Representation- based Classification (SRC) algorithm for open set recognition where not all classes presented during testing are known during training. The SRC algorithm uses class reconstruction errors for classification. As most of the discriminative information for open set recognition is hidden in the tail part of the matched and sum of non-matched reconstruction error distributions, we model the tail of those two error distributions using the statistical Extreme Value Theory (EVT). Then we simplify the open set recognition problem into a set of hypothesis testing problems. The confidence scores corresponding to the tail distributions of a novel test sample are then fused to determine its identity. The effectiveness of the proposed method is demonstrated using four publicly available image and object classification datasets and it is shown that this method can perform significantly better than many competitive open set recognition algorithms. Code is public available: https://github.com/hezhangsprinter/SROSR
Semi-supervised techniques have removed the barriers of large scale labelled set by exploiting unlabelled data to improve the performance of a model. In this paper, we propose a semi-supervised deep multi-task classification and localization approach HydraMix-Net in the field of medical imagining where labelling is time consuming and costly. Firstly, the pseudo labels are generated using the model's prediction on the augmented set of unlabelled image with averaging. The high entropy predictions are further sharpened to reduced the entropy and are then mixed with the labelled set for training. The model is trained in multi-task learning manner with noise tolerant joint loss for classification localization and achieves better performance when given limited data in contrast to a simple deep model. On DLBCL data it achieves 80\% accuracy in contrast to simple CNN achieving 70\% accuracy when given only 100 labelled examples.
Data augmentation is a commonly used approach to improving the generalization of deep learning models. Recent works show that learned data augmentation policies can achieve better generalization than hand-crafted ones. However, most of these works use unified augmentation policies for all samples in a dataset, which is observed not necessarily beneficial for all labels in multi-label classification tasks, i.e., some policies may have negative impacts on some labels while benefitting the others. To tackle this problem, we propose a novel Label-Based AutoAugmentation (LB-Aug) method for multi-label scenarios, where augmentation policies are generated with respect to labels by an augmentation-policy network. The policies are learned via reinforcement learning using policy gradient methods, providing a mapping from instance labels to their optimal augmentation policies. Numerical experiments show that our LB-Aug outperforms previous state-of-the-art augmentation methods by large margins in multiple benchmarks on image and video classification.
Recently, the Magnetic Resonance Imaging (MRI) images have limited and unsatisfactory resolutions due to various constraints such as physical, technological and economic considerations. Super-resolution techniques can obtain high-resolution MRI images. The traditional methods obtained the resolution enhancement of brain MRI by interpolations, affecting the accuracy of the following diagnose process. The requirement for brain image quality is fast increasing. In this paper, we propose an image super-resolution (SR) method based on overcomplete dictionaries and inherent similarity of an image to recover the high-resolution (HR) image from a single low-resolution (LR) image. We explore the nonlocal similarity of the image to tentatively search for similar blocks in the whole image and present a joint reconstruction method based on compressive sensing (CS) and similarity constraints. The sparsity and self-similarity of the image blocks are taken as the constraints. The proposed method is summarized in the following steps. First, a dictionary classification method based on the measurement domain is presented. The image blocks are classified into smooth, texture and edge parts by analyzing their features in the measurement domain. Then, the corresponding dictionaries are trained using the classified image blocks. Equally important, in the reconstruction part, we use the CS reconstruction method to recover the HR brain MRI image, considering both nonlocal similarity and the sparsity of an image as the constraints. This method performs better both visually and quantitatively than some existing methods.
시각 기초 모델(visual foundation model)은 이미지 및 비디오 이해의 핵심이지만, 일반적으로 대량의 데이터와 연산이 필요하다. 시각 기초 모델 사전학습에 요구되는 현재의 규모는 지속 가능하지 않거나 불필요할 수 있으며, 더 적은 자원으로 효과적인 모델을 얻을 수 있을 때 상당한 이점이 있다. 자원 제약 하에서 자기 지도 학습(SSL) 목적 함수가 어떻게 동작하는지 더 잘 이해하기 위해, 데이터·아키텍처·연산 예산을 맞춘 조건에서 이미지 및 비디오 SSL 목적 함수에 대한 통제 연구를 수행했다. 대조(contrastive), 재구(reconstruction), 특징(feature) 예측, 확산(diffusion) 목적 함수를 비교하고, 다양한 이미지·비디오 이해 과제 전반에 걸쳐 단독 및 공동 학습된 이미지-비디오 SSL 공식을 평가했다. 결과에 따르면 DINOv2 스타일 사전학습이 제한된 자원 하에서 일관되게 가장 강한 전반적 성능을 보였다. 또한 DINOv2를 VideoMAE와 같은 비디오 SSL 목적 함수와 결합하면 이미지 분류 및 분할(segmentation) 성능이 크게 향상되지만, 비디오 추적 및 카메라 포즈 추정 성능은 저하되어, 의미(semantic) 표현 학습과 기하(geometric) 표현 학습 사이의 중요한 트레이드오프가 드러났다. 이러한 결과는 자원이 제한된 설정에서 이미지와 비디오 SSL 목적 함수를 결합하는 것이 유익할 수 있음을 시사하며, 동시에 의미·시간·기하 감독을 더 잘 균형 있게 맞추는 개선된 방법의 필요성을 강조한다.
Unlike image classification and annotation, for which deep network models have achieved dominating superior performances compared to traditional computer vision algorithms, deep learning for automatic image segmentation still faces critical challenges. One of such hurdles is to obtain ground-truth segmentations as the training labels for deep network training. Especially when we study biomedical images, such as histopathological images (histo-images), it is unrealistic to ask for manual segmentation labels as the ground truth for training due to the fine image resolution as well as the large image size and complexity. In this paper, instead of relying on clean segmentation labels, we study whether and how integrating imperfect or noisy segmentation results from off-the-shelf segmentation algorithms may help achieve better segmentation results through a new Adaptive Noise-Tolerant Network (ANTN) model. We extend the noisy label deep learning to image segmentation with two novel aspects: (1) multiple noisy labels can be integrated into one deep learning model; (2) noisy segmentation modeling, including probabilistic parameters, is adaptive, depending on the given testing image appearance. Implementation of the new ANTN model on both the synthetic data and real-world histo-images demonstrates its effectiveness and superiority over off-the-shelf and other existing deep-learning-based image segmentation algorithms.
Weakly supervised object detection aims at learning precise object detectors, given image category labels. In recent prevailing works, this problem is generally formulated as a multiple instance learning module guided by an image classification loss. The object bounding box is assumed to be the one contributing most to the classification among all proposals. However, the region contributing most is also likely to be a crucial part or the supporting context of an object. To obtain a more accurate detector, in this work we propose a novel end-to-end weakly supervised detection approach, where a newly introduced generative adversarial segmentation module interacts with the conventional detection module in a collaborative loop. The collaboration mechanism takes full advantages of the complementary interpretations of the weakly supervised localization task, namely detection and segmentation tasks, forming a more comprehensive solution. Consequently, our method obtains more precise object bounding boxes, rather than parts or irrelevant surroundings. Expectedly, the proposed method achieves an accuracy of 51.0% on the PASCAL VOC 2007 dataset, outperforming the state-of-the-arts and demonstrating its superiority for weakly supervised object detection.
In this work, we present MorphoCluster, a software tool for data-driven, fast and accurate annotation of large image data sets. While already having surpassed the annotation rate of human experts, volume and complexity of marine data will continue to increase in the coming years. Still, this data requires interpretation. MorphoCluster augments the human ability to discover patterns and perform object classification in large amounts of data by embedding unsupervised clustering in an interactive process. By aggregating similar images into clusters, our novel approach to image annotation increases consistency, multiplies the throughput of an annotator and allows experts to adapt the granularity of their sorting scheme to the structure in the data. By sorting a set of 1.2M objects into 280 data-driven classes in 71 hours (16k objects per hour), with 90% of these classes having a precision of 0.889 or higher. This shows that MorphoCluster is at the same time fast, accurate and consistent, provides a fine-grained and data-driven classification and enables novelty detection. MorphoCluster is available as open-source software at https://github.com/morphocluster.
We demonstrate that a deep model can be reparametrized as a linear combination of several randomly initialized and frozen deep models in the weight space. During training, we seek local minima that reside within the subspace spanned by these random models (i.e., `basis' networks). Our framework, PRANC, enables significant compaction of a deep model. The model can be reconstructed using a single scalar `seed,' employed to generate the pseudo-random `basis' networks, together with the learned linear mixture coefficients. In practical applications, PRANC addresses the challenge of efficiently storing and communicating deep models, a common bottleneck in several scenarios, including multi-agent learning, continual learners, federated systems, and edge devices, among others. In this study, we employ PRANC to condense image classification models and compress images by compacting their associated implicit neural networks. PRANC outperforms baselines with a large margin on image classification when compressing a deep model almost $100$ times. Moreover, we show that PRANC enables memory-efficient inference by generating layer-wise weights on the fly. The source code of PRANC is here: \url{https://github.com/UCDvision/PRANC}
Jewellery item retrieval is regularly used to find what people want on online marketplaces using a sample query reference image. Considering recent developments, due to the simultaneous nature of various jewelry items, various jewelry goods' occlusion in images or visual streams, as well as shape deformation, content-based jewellery item retrieval (CBJIR) still has limitations whenever it pertains to visual searching in the actual world. This article proposed a content-based jewellery item retrieval method using the local region-based histograms in HSV color space. Using five local regions, our novel jewellery classification module extracts the specific feature vectors from the query image. The jewellery classification module is also applied to the jewellery database to extract feature vectors. Finally, the similarity score is matched between the database and query features vectors to retrieve the jewellery items from the database. The proposed method performance is tested on publicly available jewellery item retrieval datasets, i.e. ringFIR and Fashion Product Images dataset. The experimental results demonstrate the dominance of the proposed method over the baseline methods for retrieving desired jewellery products.
This study compared the classification performance of Gemini Pro and GPT-4V in educational settings. Employing visual question answering (VQA) techniques, the study examined both models' abilities to read text-based rubrics and then automatically score student-drawn models in science education. We employed both quantitative and qualitative analyses using a dataset derived from student-drawn scientific models and employing NERIF (Notation-Enhanced Rubrics for Image Feedback) prompting methods. The findings reveal that GPT-4V significantly outperforms Gemini Pro in terms of scoring accuracy and Quadratic Weighted Kappa. The qualitative analysis reveals that the differences may be due to the models' ability to process fine-grained texts in images and overall image classification performance. Even adapting the NERIF approach by further de-sizing the input images, Gemini Pro seems not able to perform as well as GPT-4V. The findings suggest GPT-4V's superior capability in handling complex multimodal educational tasks. The study concludes that while both models represent advancements in AI, GPT-4V's higher performance makes it a more suitable tool for educational applications involving multimodal data interpretation.
Model merging is a technique that combines multiple finetuned models into a single model without additional training, allowing a free-rider to cheaply inherit specialized capabilities. This study investigates methodologies to suppress unwanted model merging by free-riders. Existing methods such as model watermarking or fingerprinting can only detect merging in hindsight. In contrast, we propose a first proactive defense against model merging. Specifically, our defense method modifies the model parameters so that the model is disrupted if the model is merged with any other model, while its functionality is kept unchanged if not merged with others. Our approach consists of two modules, rearranging MLP parameters and scaling attention heads, which push the model out of the shared basin in parameter space, causing the merging performance with other models to degrade significantly. We conduct extensive experiments on image classification, image generation, and text classification to demonstrate that our defense severely disrupts merging while retaining the functionality of the post-protect model. Moreover, we analyze potential adaptive attacks and further propose a dropout-based pruning to improve our proposal's robustness.