Can we train a hybrid discriminative-generative model within a single network? This question has recently been answered in the affirmative, introducing the field of Joint Energy-based Model (JEM), which achieves high classification accuracy and image generation quality simultaneously. Despite recent advances, there remain two performance gaps: the accuracy gap to the standard softmax classifier, and the generation quality gap to state-of-the-art generative models. In this paper, we introduce a variety of training techniques to bridge the accuracy gap and the generation quality gap of JEM. 1) We incorporate a recently proposed sharpness-aware minimization (SAM) framework to train JEM, which promotes the energy landscape smoothness and the generalizability of JEM. 2) We exclude data augmentation from the maximum likelihood estimate pipeline of JEM, and mitigate the negative impact of data augmentation to image generation quality. Extensive experiments on multiple datasets demonstrate that our SADA-JEM achieves state-of-the-art performances and outperforms JEM in image classification, image generation, calibration, out-of-distribution detection and adversarial robustness by a notable margin. Our code is available at https://github.com/sndnyang/SADAJEM.
Active Learning (AL) is a user-interactive approach aimed at reducing annotation costs by selecting the most crucial examples to label. Although AL has been extensively studied for image classification tasks, the specific scenario of interactive image retrieval has received relatively little attention. This scenario presents unique characteristics, including an open-set and class-imbalanced binary classification, starting with very few labeled samples. We introduce a novel batch-mode Active Learning framework named GAL (Greedy Active Learning) that better copes with this application. It incorporates a new acquisition function for sample selection that measures the impact of each unlabeled sample on the classifier. We further embed this strategy in a greedy selection approach, better exploiting the samples within each batch. We evaluate our framework with both linear (SVM) and non-linear MLP/Gaussian Process classifiers. For the Gaussian Process case, we show a theoretical guarantee on the greedy approximation. Finally, we assess our performance for the interactive content-based image retrieval task on several benchmarks and demonstrate its superiority over existing approaches and common baselines. Code is available at https://github.com/barleah/GreedyAL.
The main challenge for fine-grained few-shot image classification is to learn feature representations with higher inter-class and lower intra-class variations, with a mere few labelled samples. Conventional few-shot learning methods however cannot be naively adopted for this fine-grained setting -- a quick pilot study reveals that they in fact push for the opposite (i.e., lower inter-class variations and higher intra-class variations). To alleviate this problem, prior works predominately use a support set to reconstruct the query image and then utilize metric learning to determine its category. Upon careful inspection, we further reveal that such unidirectional reconstruction methods only help to increase inter-class variations and are not effective in tackling intra-class variations. In this paper, we for the first time introduce a bi-reconstruction mechanism that can simultaneously accommodate for inter-class and intra-class variations. In addition to using the support set to reconstruct the query set for increasing inter-class variations, we further use the query set to reconstruct the support set for reducing intra-class variations. This design effectively helps the model to explore more subtle and discriminative features which is key for the fine-grained problem in hand. Furthermore, we also construct a self-reconstruction module to work alongside the bi-directional module to make the features even more discriminative. Experimental results on three widely used fine-grained image classification datasets consistently show considerable improvements compared with other methods. Codes are available at: https://github.com/PRIS-CV/Bi-FRN.
The availability of the sheer volume of Copernicus Sentinel-2 imagery has created new opportunities for exploiting deep learning (DL) methods for land use land cover (LULC) image classification. However, an extensive set of benchmark experiments is currently lacking, i.e. DL models tested on the same dataset, with a common and consistent set of metrics, and in the same hardware. In this work, we use the BigEarthNet Sentinel-2 dataset to benchmark for the first time different state-of-the-art DL models for the multi-label, multi-class LULC image classification problem, contributing with an exhaustive zoo of 60 trained models. Our benchmark includes standard CNNs, as well as non-convolutional methods. We put to the test EfficientNets and Wide Residual Networks (WRN) architectures, and leverage classification accuracy, training time and inference rate. Furthermore, we propose to use the EfficientNet framework for the compound scaling of a lightweight WRN. Enhanced with an Efficient Channel Attention mechanism, our scaled lightweight model emerged as the new state-of-the-art. It achieves 4.5% higher averaged F-Score classification accuracy for all 19 LULC classes compared to a standard ResNet50 baseline model, with an order of magnitude less trainable parameters. We provide access to all trained models, along with our code for distributed training on multiple GPU nodes. This model zoo of pre-trained encoders can be used for transfer learning and rapid prototyping in different remote sensing tasks that use Sentinel-2 data, instead of exploiting backbone models trained with data from a different domain, e.g., from ImageNet. We validate their suitability for transfer learning in different datasets of diverse volumes. Our top-performing WRN achieves state-of-the-art performance (71.1% F-Score) on the SEN12MS dataset while being exposed to only a small fraction of the training dataset.
In this paper, we explore and compare multiple solutions to the problem of data augmentation in image classification. Previous work has demonstrated the effectiveness of data augmentation through simple techniques, such as cropping, rotating, and flipping input images. We artificially constrain our access to data to a small subset of the ImageNet dataset, and compare each data augmentation technique in turn. One of the more successful data augmentations strategies is the traditional transformations mentioned above. We also experiment with GANs to generate images of different styles. Finally, we propose a method to allow a neural net to learn augmentations that best improve the classifier, which we call neural augmentation. We discuss the successes and shortcomings of this method on various datasets.
Convolutional Neural Networks (CNNs) have been the standard for image classification tasks for a long time, but more recently attention-based mechanisms have gained traction. This project aims to compare traditional CNNs with attention-augmented CNNs across an image classification task. By evaluating and comparing their performance, accuracy and computational efficiency, the project will highlight benefits and trade-off of the localized feature extraction of traditional CNNs and the global context capture in attention-augmented CNNs. By doing this, we can reveal further insights into their respective strengths and weaknesses, guide the selection of models based on specific application needs and ultimately, enhance understanding of these architectures in the deep learning community. This was our final project for CS7643 Deep Learning course at Georgia Tech.
Fine-grained image classification is to recognize hundreds of subcategories belonging to the same basic-level category, such as 200 subcategories belonging to the bird, which is highly challenging due to large variance in the same subcategory and small variance among different subcategories. Existing methods generally first locate the objects or parts and then discriminate which subcategory the image belongs to. However, they mainly have two limitations: (1) Relying on object or part annotations which are heavily labor consuming. (2) Ignoring the spatial relationships between the object and its parts as well as among these parts, both of which are significantly helpful for finding discriminative parts. Therefore, this paper proposes the object-part attention model (OPAM) for weakly supervised fine-grained image classification, and the main novelties are: (1) Object-part attention model integrates two level attentions: object-level attention localizes objects of images, and part-level attention selects discriminative parts of object. Both are jointly employed to learn multi-view and multi-scale features to enhance their mutual promotions. (2) Object-part spatial constraint model combines two spatial constraints: object spatial constraint ensures selected parts highly representative, and part spatial constraint eliminates redundancy and enhances discrimination of selected parts. Both are jointly employed to exploit the subtle and local differences for distinguishing the subcategories. Importantly, neither object nor part annotations are used in our proposed approach, which avoids the heavy labor consumption of labeling. Comparing with more than 10 state-of-the-art methods on 4 widely-used datasets, our OPAM approach achieves the best performance.
Large-scale image databases such as ImageNet have significantly advanced image classification and other visual recognition tasks. However much of these datasets are constructed only for single-label and coarse object-level classification. For real-world applications, multiple labels and fine-grained categories are often needed, yet very few such datasets exist publicly, especially those of large-scale and high quality. In this work, we contribute to the community a new dataset called iMaterialist Fashion Attribute (iFashion-Attribute) to address this problem in the fashion domain. The dataset is constructed from over one million fashion images with a label space that includes 8 groups of 228 fine-grained attributes in total. Each image is annotated by experts with multiple, high-quality fashion attributes. The result is the first known million-scale multi-label and fine-grained image dataset. We conduct extensive experiments and provide baseline results with modern deep Convolutional Neural Networks (CNNs). Additionally, we demonstrate models pre-trained on iFashion-Attribute achieve superior transfer learning performance on fashion related tasks compared with pre-training from ImageNet or other fashion datasets. Data is available at: https://github.com/visipedia/imat_fashion_comp
In some medical imaging tasks and other settings where only small parts of the image are informative for the classification task, traditional CNNs can sometimes struggle to generalise. Manually annotated Regions of Interest (ROI) are sometimes used to isolate the most informative parts of the image. However, these are expensive to collect and may vary significantly across annotators. To overcome these issues, we propose a framework that employs saliency maps to obtain soft spatial attention masks that modulate the image features at different scales. We refer to our method as Adversarial Counterfactual Attention (ACAT). ACAT increases the baseline classification accuracy of lesions in brain CT scans from 71.39% to 72.55% and of COVID-19 related findings in lung CT scans from 67.71% to 70.84% and exceeds the performance of competing methods. We investigate the best way to generate the saliency maps employed in our architecture and propose a way to obtain them from adversarially generated counterfactual images. They are able to isolate the area of interest in brain and lung CT scans without using any manual annotations. In the task of localising the lesion location out of 6 possible regions, they obtain a score of 65.05% on brain CT scans, improving the score of 61.29% obtained with the best competing method.
Knowledge is acquired by humans through experience, and no boundary is set between the kinds of knowledge or skill levels we can achieve on different tasks at the same time. When it comes to Neural Networks, that is not the case. The breakthroughs in the field are extremely task and domain-specific. Vision and language are dealt with in separate manners, using separate methods and different datasets. Current text classification methods, mostly rely on obtaining contextual embeddings for input text samples, then training a classifier on the embedded dataset. Transfer learning in Language-related tasks in general, is heavily used in obtaining the contextual text embeddings for the input samples. In this work, we propose to use the knowledge acquired by benchmark Vision Models which are trained on ImageNet to help a much smaller architecture learn to classify text. A data transformation technique is used to create a new image dataset, where each image represents a sentence embedding from the last six layers of BERT, projected on a 2D plane using a t-SNE based method. We trained five models containing early layers sliced from vision models which are pretrained on ImageNet, on the created image dataset for the IMDB dataset embedded with the last six layers of BERT. Despite the challenges posed by the very different datasets, experimental results achieved by this approach which links large pretrained models on both language and vision, are very promising, without employing compute resources. Specifically, Sentiment Analysis is achieved by five different models on the same image dataset obtained after BERT embeddings are transformed into gray scale images. Index Terms: BERT, Convolutional Neural Networks, Domain Adaptation, image classification, Natural Language Processing, t-SNE, text classification, Transfer Learning
Hyperspectral Image (HSI) classification using Convolutional Neural Networks (CNN) is widely found in the current literature. Approaches vary from using SVMs to 2D CNNs, 3D CNNs, 3D-2D CNNs. Besides 3D-2D CNNs and FuSENet, the other approaches do not consider both the spectral and spatial features together for HSI classification task, thereby resulting in poor performances. 3D CNNs are computationally heavy and are not widely used, while 2D CNNs do not consider multi-resolution processing of images, and only limits itself to the spatial features. Even though 3D-2D CNNs try to model the spectral and spatial features their performance seems limited when applied over multiple dataset. In this article, we propose SpectralNET, a wavelet CNN, which is a variation of 2D CNN for multi-resolution HSI classification. A wavelet CNN uses layers of wavelet transform to bring out spectral features. Computing a wavelet transform is lighter than computing 3D CNN. The spectral features extracted are then connected to the 2D CNN which bring out the spatial features, thereby creating a spatial-spectral feature vector for classification. Overall a better model is achieved that can classify multi-resolution HSI data with high accuracy. Experiments performed with SpectralNET on benchmark dataset, i.e. Indian Pines, University of Pavia, and Salinas Scenes confirm the superiority of proposed SpectralNET with respect to the state-of-the-art methods. The code is publicly available in https://github.com/tanmay-ty/SpectralNET.
Convolutional Neural Networks (CNN) have become state-of-the-art in the field of image classification. However, not everything is understood about their inner representations. This paper tackles the interpretability and explainability of the predictions of CNNs for multi-class classification problems. Specifically, we propose a novel visualization method of pixel-wise input attribution called Softmax-Gradient Layer-wise Relevance Propagation (SGLRP). The proposed model is a class discriminate extension to Deep Taylor Decomposition (DTD) using the gradient of softmax to back propagate the relevance of the output probability to the input image. Through qualitative and quantitative analysis, we demonstrate that SGLRP can successfully localize and attribute the regions on input images which contribute to a target object's classification. We show that the proposed method excels at discriminating the target objects class from the other possible objects in the images. We confirm that SGLRP performs better than existing Layer-wise Relevance Propagation (LRP) based methods and can help in the understanding of the decision process of CNNs.
Extracting discriminative features plays a crucial role in the fine-grained visual classification task. Most of the existing methods focus on developing attention or augmentation mechanisms to achieve this goal. However, addressing the ambiguity in the top-k prediction classes is not fully investigated. In this paper, we introduce a Self Assessment Classifier, which simultaneously leverages the representation of the image and top-k prediction classes to reassess the classification results. Our method is inspired by continual learning with coarse-grained and fine-grained classifiers to increase the discrimination of features in the backbone and produce attention maps of informative areas on the image. In practice, our method works as an auxiliary branch and can be easily integrated into different architectures. We show that by effectively addressing the ambiguity in the top-k prediction classes, our method achieves new state-of-the-art results on CUB200-2011, Stanford Dog, and FGVC Aircraft datasets. Furthermore, our method also consistently improves the accuracy of different existing fine-grained classifiers with a unified setup.
A major challenge in matching images and text is that they have intrinsically different data distributions and feature representations. Most existing approaches are based either on embedding or classification, the first one mapping image and text instances into a common embedding space for distance measuring, and the second one regarding image-text matching as a binary classification problem. Neither of these approaches can, however, balance the matching accuracy and model complexity well. We propose a novel framework that achieves remarkable matching performance with acceptable model complexity. Specifically, in the training stage, we propose a novel Multi-modal Tensor Fusion Network (MTFN) to explicitly learn an accurate image-text similarity function with rank-based tensor fusion rather than seeking a common embedding space for each image-text instance. Then, during testing, we deploy a generic Cross-modal Re-ranking (RR) scheme for refinement without requiring additional training procedure. Extensive experiments on two datasets demonstrate that our MTFN-RR consistently achieves the state-of-the-art matching performance with much less time complexity. The implementation code is available at https://github.com/Wangt-CN/MTFN-RR-PyTorch-Code.
Can we leverage high-resolution information without the unsustainable quadratic complexity to input scale? We propose Traversal Network (TNet), a novel multi-scale hard-attention architecture, which traverses image scale-space in a top-down fashion, visiting only the most informative image regions along the way. TNet offers an adjustable trade-off between accuracy and complexity, by changing the number of attended image locations. We compare our model against hard-attention baselines on ImageNet, achieving higher accuracy with less resources (FLOPs, processing time and memory). We further test our model on fMoW dataset, where we process satellite images of size up to $896 \times 896$ px, getting up to $2.5$x faster processing compared to baselines operating on the same resolution, while achieving higher accuracy as well. TNet is modular, meaning that most classification models could be adopted as its backbone for feature extraction, making the reported performance gains orthogonal to benefits offered by existing optimized deep models. Finally, hard-attention guarantees a degree of interpretability to our model's predictions, without any extra cost beyond inference. Code is available at $\href{https://github.com/Tpap/TNet}{github.com/Tpap/TNet}$.
In this paper, we present CLCC, a novel contrastive learning framework for color constancy. Contrastive learning has been applied for learning high-quality visual representations for image classification. One key aspect to yield useful representations for image classification is to design illuminant invariant augmentations. However, the illuminant invariant assumption conflicts with the nature of the color constancy task, which aims to estimate the illuminant given a raw image. Therefore, we construct effective contrastive pairs for learning better illuminant-dependent features via a novel raw-domain color augmentation. On the NUS-8 dataset, our method provides $17.5\%$ relative improvements over a strong baseline, reaching state-of-the-art performance without increasing model complexity. Furthermore, our method achieves competitive performance on the Gehler dataset with $3\times$ fewer parameters compared to top-ranking deep learning methods. More importantly, we show that our model is more robust to different scenes under close proximity of illuminants, significantly reducing $28.7\%$ worst-case error in data-sparse regions.
Feature preference in Convolutional Neural Network (CNN) image classifiers is integral to their decision making process, and while the topic has been well studied, it is still not understood at a fundamental level. We test a range of task relevant feature attributes (including shape, texture, and color) with varying degrees of signal and noise in highly controlled CNN image classification experiments using synthetic datasets to determine feature preferences. We find that CNNs will prefer features with stronger signal strength and lower noise irrespective of whether the feature is texture, shape, or color. This provides guidance for a predictive model for task relevant feature preferences, demonstrates pathways for bias in machine models that can be avoided with careful controls on experimental setup, and suggests that comparisons between how humans and machines prefer task relevant features in vision classification tasks should be revisited. Code to reproduce experiments in this paper can be found at \url{https://github.com/mwolff31/signal_preference}.
Multi-modal models require aligned, shared embedding spaces. However, common CLIP-based approaches need large amounts of samples and do not natively support 3D or tabular data, both of which are crucial in the medical domain. To address these issues, we revisit CLIP-style alignment by training a domain-specific 3D foundation model as an image encoder and demonstrate that modality alignment is feasible with only 62 MRI scans. Our approach is enabled by a simple embedding accumulation strategy required for training in 3D, which scales the amount of negative pairs across batches in order to stabilize training. We perform a thorough evaluation of various design choices, including the choice of backbone and loss functions, and evaluate the proposed methodology on zero-shot classification and image-retrieval tasks. While zero-shot image-retrieval remains challenging, zero-shot classification results demonstrate that the proposed approach can meaningfully align the representations of 3D MRI with tabular data.
Label information plays an important role in supervised hyperspectral image classification problem. However, current classification methods all ignore an important and inevitable problem---labels may be corrupted and collecting clean labels for training samples is difficult, and often impractical. Therefore, how to learn from the database with noisy labels is a problem of great practical importance. In this paper, we study the influence of label noise on hyperspectral image classification, and develop a random label propagation algorithm (RLPA) to cleanse the label noise. The key idea of RLPA is to exploit knowledge (e.g., the superpixel based spectral-spatial constraints) from the observed hyperspectral images and apply it to the process of label propagation. Specifically, RLPA first constructs a spectral-spatial probability transfer matrix (SSPTM) that simultaneously considers the spectral similarity and superpixel based spatial information. It then randomly chooses some training samples as "clean" samples and sets the rest as unlabeled samples, and propagates the label information from the "clean" samples to the rest unlabeled samples with the SSPTM. By repeating the random assignment (of "clean" labeled samples and unlabeled samples) and propagation, we can obtain multiple labels for each training sample. Therefore, the final propagated label can be calculated by a majority vote algorithm. Experimental studies show that RLPA can reduce the level of noisy label and demonstrates the advantages of our proposed method over four major classifiers with a significant margin---the gains in terms of the average OA, AA, Kappa are impressive, e.g., 9.18%, 9.58%, and 0.1043. The Matlab source code is available at https://github.com/junjun-jiang/RLPA
Recently, the Vision Transformer (ViT) model has replaced the classical Convolutional Neural Network (ConvNet) in various computer vision tasks due to its superior performance. Even in hyperspectral image (HSI) classification field, ViT-based methods also show promising potential. Nevertheless, ViT encounters notable difficulties in processing HSI data. Its self-attention mechanism, which exhibits quadratic complexity, escalates computational costs. Additionally, ViT's substantial demand for training samples does not align with the practical constraints posed by the expensive labeling of HSI data. To overcome these challenges, we propose a 3D relational ConvNet named 3D-RCNet, which inherits both strengths of ConvNet and ViT, resulting in high performance in HSI classification. We embed the self-attention mechanism of Transformer into the convolutional operation of ConvNet to design 3D relational convolutional operation and use it to build the final 3D-RCNet. The proposed 3D-RCNet maintains the high computational efficiency of ConvNet while enjoying the flexibility of ViT. Additionally, the proposed 3D relational convolutional operation is a plug-and-play operation, which can be inserted into previous ConvNet-based HSI classification methods seamlessly. Empirical evaluations on three representative benchmark HSI datasets show that the proposed model outperforms previous ConvNet-based and ViT-based HSI approaches.