Search

The fusion of vision and language has brought about a transformative shift in computer vision through the emergence of Vision-Language Models (VLMs). However, the resource-intensive nature of existing VLMs poses a significant challenge. We need an accessible method for developing the next generation of VLMs. To address this issue, we propose Zoom-shot, a novel method for transferring the zero-shot capabilities of CLIP to any pre-trained vision encoder. We do this by exploiting the multimodal information (i.e. text and image) present in the CLIP latent space through the use of specifically designed multimodal loss functions. These loss functions are (1) cycle-consistency loss and (2) our novel prompt-guided knowledge distillation loss (PG-KD). PG-KD combines the concept of knowledge distillation with CLIP's zero-shot classification, to capture the interactions between text and image features. With our multimodal losses, we train a $\textbf{linear mapping}$ between the CLIP latent space and the latent space of a pre-trained vision encoder, for only a $\textbf{single epoch}$. Furthermore, Zoom-shot is entirely unsupervised and is trained using $\textbf{unpaired}$ data. We test the zero-shot capabilities of a range of vision encoders augmented as new VLMs, on coarse and fine-grained classification datasets, outperforming the previous state-of-the-art in this problem domain. In our ablations, we find Zoom-shot allows for a trade-off between data and compute during training; and our state-of-the-art results can be obtained by reducing training from 20% to 1% of the ImageNet training data with 20 epochs. All code and models are available on GitHub., Comment: 15 pages

In this work, we investigate the problem of Model-Agnostic Zero-Shot Classification (MA-ZSC), which refers to training non-specific classification architectures (downstream models) to classify real images without using any real images during training. Recent research has demonstrated that generating synthetic training images using diffusion models provides a potential solution to address MA-ZSC. However, the performance of this approach currently falls short of that achieved by large-scale vision-language models. One possible explanation is a potential significant domain gap between synthetic and real images. Our work offers a fresh perspective on the problem by providing initial insights that MA-ZSC performance can be improved by improving the diversity of images in the generated dataset. We propose a set of modifications to the text-to-image generation process using a pre-trained diffusion model to enhance diversity, which we refer to as our $\textbf{bag of tricks}$. Our approach shows notable improvements in various classification architectures, with results comparable to state-of-the-art models such as CLIP. To validate our approach, we conduct experiments on CIFAR10, CIFAR100, and EuroSAT, which is particularly difficult for zero-shot classification due to its satellite image domain. We evaluate our approach with five classification architectures, including ResNet and ViT. Our findings provide initial insights into the problem of MA-ZSC using diffusion models. All code will be available on GitHub., Comment: (10 pages, 6 figures, 3 tables, preprint)

The Once-For-All (OFA) method offers an excellent pathway to deploy a trained neural network model into multiple target platforms by utilising the supernet-subnet architecture. Once trained, a subnet can be derived from the supernet (both architecture and trained weights) and deployed directly to the target platform with little to no retraining or fine-tuning. To train the subnet population, OFA uses a novel training method called Progressive Shrinking (PS) which is designed to limit the negative impact of interference during training. It is believed that higher interference during training results in lower subnet population accuracies. In this work we take a second look at this interference effect. Surprisingly, we find that interference mitigation strategies do not have a large impact on the overall subnet population performance. Instead, we find the subnet architecture selection bias during training to be a more important aspect. To show this, we propose a simple-yet-effective method called Random Subnet Sampling (RSS), which does not have mitigation on the interference effect. Despite no mitigation, RSS is able to produce a better performing subnet population than PS in four small-to-medium-sized datasets; suggesting that the interference effect does not play a pivotal role in these datasets. Due to its simplicity, RSS provides a $1.9\times$ reduction in training times compared to PS. A $6.1\times$ reduction can also be achieved with a reasonable drop in performance when the number of RSS training epochs are reduced. Code available at https://github.com/Jordan-HS/RSS-Interference-CVPRW2022., Comment: Accepted to CVPR Embedded Vision Workshop 2022

We present a novel approach to perform the unsupervised domain adaptation for object detection through forward-backward cyclic (FBC) training. Recent adversarial training based domain adaptation methods have shown their effectiveness on minimizing domain discrepancy via marginal feature distributions alignment. However, aligning the marginal feature distributions does not guarantee the alignment of class conditional distributions. This limitation is more evident when adapting object detectors as the domain discrepancy is larger compared to the image classification task, e.g. various number of objects exist in one image and the majority of content in an image is the background. This motivates us to learn domain invariance for category level semantics via gradient alignment. Intuitively, if the gradients of two domains point in similar directions, then the learning of one domain can improve that of another domain. To achieve gradient alignment, we propose Forward-Backward Cyclic Adaptation, which iteratively computes adaptation from source to target via backward hopping and from target to source via forward passing. In addition, we align low-level features for adapting holistic color/texture via adversarial training. However, the detector performs well on both domains is not ideal for target domain. As such, in each cycle, domain diversity is enforced by maximum entropy regularization on the source domain to penalize confident source-specific learning and minimum entropy regularization on target domain to intrigue target-specific learning. Theoretical analysis of the training process is provided, and extensive experiments on challenging cross-domain object detection datasets have shown the superiority of our approach over the state-of-the-art.

The condition of the Glomeruli, or filter sacks, in renal Direct Immunofluorescence (DIF) specimens is a critical indicator for diagnosing kidney diseases. A digital pathology system which digitizes a glass histology slide into a Whole Slide Image (WSI) and then automatically detects and zooms in on the glomeruli with a higher magnification objective will be extremely helpful for pathologists. In this paper, using glomerulus detection as the study case, we provide analysis and observations on several important issues to help with the development of Computer Aided Diagnostic (CAD) systems to process WSIs. Large image resolution, large file size, and data scarcity are always challenging to deal with. To this end, we first examine image downsampling rates in terms of their effect on detection accuracy. Second, we examine the impact of image compression. Third, we examine the relationship between the size of the training set and detection accuracy. To understand the above issues, experiments are performed on the state-of-the-art detectors: Faster R-CNN, R-FCN, Mask R-CNN and SSD. Critical findings are observed: (1) The best balance between detection accuracy, detection speed and file size is achieved at 8 times downsampling captured with a $40\times$ objective; (2) compression which reduces the file size dramatically, does not necessarily have an adverse effect on overall accuracy; (3) reducing the amount of training data to some extents causes a drop in precision but has a negligible impact on the recall; (4) in most cases, Faster R-CNN achieves the best accuracy in the glomerulus detection task. We show that the image file size of $40\times$ WSI images can be reduced by a factor of over 6000 with negligible loss of glomerulus detection accuracy.

Electrical distribution poles are important assets in electricity supply. These poles need to be maintained in good condition to ensure they protect community safety, maintain reliability of supply, and meet legislative obligations. However, maintaining such a large volumes of assets is an expensive and challenging task. To address this, recent approaches utilise imagery data captured from helicopter and/or drone inspections. Whilst reducing the cost for manual inspection, manual analysis on each image is still required. As such, several image-based automated inspection systems have been proposed. In this paper, we target two major challenges: tiny object detection and extremely imbalanced datasets, which currently hinder the wide deployment of the automatic inspection. We propose a novel two-stage zoom-in detection method to gradually focus on the object of interest. To address the imbalanced dataset problem, we propose the resampling as well as reweighting schemes to iteratively adapt the model to the large intra-class variation of major class and balance the contributions to the loss from each class. Finally, we integrate these components together and devise a novel automatic inspection framework. Extensive experiments demonstrate that our proposed approaches are effective and can boost the performance compared to the baseline methods., Comment: electrical distribution pole inspection, integrated inspection system, object detection, imbalanced data classification, To appear in Proceeding of ICIP 2019

Histopathology image analysis can be considered as a Multiple instance learning (MIL) problem, where the whole slide histopathology image (WSI) is regarded as a bag of instances (i.e, patches) and the task is to predict a single class label to the WSI. However, in many real-life applications such as computational pathology, discovering the key instances that trigger the bag label is of great interest because it provides reasons for the decision made by the system. In this paper, we propose a deep convolutional neural network (CNN) model that addresses the primary task of a bag classification on a WSI and also learns to identify the response of each instance to provide interpretable results to the final prediction. We incorporate the attention mechanism into the proposed model to operate the transformation of instances and learn attention weights to allow us to find key patches. To perform a balanced training, we introduce adaptive weighing in each training bag to explicitly adjust the weight distribution in order to concentrate more on the contribution of hard samples. Based on the learned attention weights, we further develop a solution to boost the classification performance by generating the bags with hard negative instances. We conduct extensive experiments on colon and breast cancer histopathology data and show that our framework achieves state-of-the-art performance., Comment: Accepted by MICCAI 2019

This work tackles the problem of generating a medical report for multi-image panels. We apply our solution to the Renal Direct Immunofluorescence (RDIF) assay which requires a pathologist to generate a report based on observations across the eight different WSI in concert with existing clinical features. To this end, we propose a novel attention-based multi-modal generative recurrent neural network (RNN) architecture capable of dynamically sampling image data concurrently across the RDIF panel. The proposed methodology incorporates text from the clinical notes of the requesting physician to regulate the output of the network to align with the overall clinical context. In addition, we found the importance of regularizing the attention weights for word generation processes. This is because the system can ignore the attention mechanism by assigning equal weights for all members. Thus, we propose two regularizations which force the system to utilize the attention mechanism. Experiments on our novel collection of RDIF WSIs provided by a large clinical laboratory demonstrate that our framework offers significant improvements over existing methods., Comment: Accepted for MICCAI 2019

The effectiveness of Symmetric Positive Definite (SPD) manifold features has been proven in various computer vision tasks. However, due to the non-Euclidean geometry of these features, existing Euclidean machineries cannot be directly used. In this paper, we tackle the classification tasks with limited training data on SPD manifolds. Our proposed framework, named Manifold Convex Class Model, represents each class on SPD manifolds using a convex model, and classification can be performed by computing distances to the convex models. We provide three methods based on different metrics to address the optimization problem of the smallest distance of a point to the convex model on SPD manifold. The efficacy of our proposed framework is demonstrated both on synthetic data and several computer vision tasks including object recognition, texture classification, person re-identification and traffic scene classification.

This work tackles the automatic fine-grained slide quality assessment problem for digitized direct smears test using the Gram staining protocol. Automatic quality assessment can provide useful information for the pathologists and the whole digital pathology workflow. For instance, if the system found a slide to have a low staining quality, it could send a request to the automatic slide preparation system to remake the slide. If the system detects severe damage in the slides, it could notify the experts that manual microscope reading may be required. In order to address the quality assessment problem, we propose a deep neural network based framework to automatically assess the slide quality in a semantic way. Specifically, the first step of our framework is to perform dense fine-grained region classification on the whole slide and calculate the region distribution histogram. Next, our framework will generate assessments of the slide quality from various perspectives: staining quality, information density, damage level and which regions are more valuable for subsequent high-magnification analysis. To make the information more accessible, we present our results in the form of a heat map and text summaries. Additionally, in order to stimulate research in this direction, we propose a novel dataset for slide quality assessment. Experiments show that the proposed framework outperforms recent related works.

This work shows that it is possible to fool/attack recent state-of-the-art face detectors which are based on the single-stage networks. Successfully attacking face detectors could be a serious malware vulnerability when deploying a smart surveillance system utilizing face detectors. We show that existing adversarial perturbation methods are not effective to perform such an attack, especially when there are multiple faces in the input image. This is because the adversarial perturbation specifically generated for one face may disrupt the adversarial perturbation for another face. In this paper, we call this problem the Instance Perturbation Interference (IPI) problem. This IPI problem is addressed by studying the relationship between the deep neural network receptive field and the adversarial perturbation. As such, we propose the Localized Instance Perturbation (LIP) that uses adversarial perturbation constrained to the Effective Receptive Field (ERF) of a target to perform the attack. Experiment results show the LIP method massively outperforms existing adversarial perturbation generation methods -- often by a factor of 2 to 10., Comment: to appear ECCV 2018 (accepted version)

This work tackles the face recognition task on images captured using thermal camera sensors which can operate in the non-light environment. While it can greatly increase the scope and benefits of the current security surveillance systems, performing such a task using thermal images is a challenging problem compared to face recognition task in the Visible Light Domain (VLD). This is partly due to the much smaller amount number of thermal imagery data collected compared to the VLD data. Unfortunately, direct application of the existing very strong face recognition models trained using VLD data into the thermal imagery data will not produce a satisfactory performance. This is due to the existence of the domain gap between the thermal and VLD images. To this end, we propose a Thermal-to-Visible Generative Adversarial Network (TV-GAN) that is able to transform thermal face images into their corresponding VLD images whilst maintaining identity information which is sufficient enough for the existing VLD face recognition models to perform recognition. Some examples are presented in Figure 1. Unlike the previous methods, our proposed TV-GAN uses an explicit closed-set face recognition loss to regularize the discriminator network training. This information will then be conveyed into the generator network in the forms of gradient loss. In the experiment, we show that by using this additional explicit regularization for the discriminator network, the TV-GAN is able to preserve more identity information when translating a thermal image of a person which is not seen before by the TV-GAN.

Automatic attribute discovery methods have gained in popularity to extract sets of visual attributes from images or videos for various tasks. Despite their good performance in some classification tasks, it is difficult to evaluate whether the attributes discovered by these methods are meaningful and which methods are the most appropriate to discover attributes for visual descriptions. In its simplest form, such an evaluation can be performed by manually verifying whether there is any consistent identifiable visual concept distinguishing between positive and negative exemplars labelled by an attribute. This manual checking is tedious, expensive and labour intensive. In addition, comparisons between different methods could also be problematic as it is not clear how one could quantitatively decide which attribute is more meaningful than the others. In this paper, we propose a novel attribute meaningfulness metric to address this challenging problem. With this metric, automatic quantitative evaluation can be performed on the attribute sets; thus, reducing the enormous effort to perform manual evaluation. The proposed metric is applied to some recent automatic attribute discovery and hashing methods on four attribute-labelled datasets. To further validate the efficacy of the proposed method, we conducted a user study. In addition, we also compared our metric with a semi-supervised attribute discover method using the mixture of probabilistic PCA. In our evaluation, we gleaned several insights that could be beneficial in developing new automatic attribute discovery methods., Comment: Submission to Pattern Recognition

Can we predict the winner of Miss Universe after watching how they stride down the catwalk during the evening gown competition? Fashion gurus say they can! In our work, we study this question from the perspective of computer vision. In particular, we want to understand whether existing computer vision approaches can be used to automatically extract the qualities exhibited by the Miss Universe winners during their catwalk. This study can pave the way towards new vision-based applications for the fashion industry. To this end, we propose a novel video dataset, called the Miss Universe dataset, comprising 10 years of the evening gown competition selected between 1996-2010. We further propose two ranking-related problems: (1) Miss Universe Listwise Ranking and (2) Miss Universe Pairwise Ranking. In addition, we also develop an approach that simultaneously addresses the two proposed problems. To describe the videos we employ the recently proposed Stacked Fisher Vectors in conjunction with robust local spatio-temporal features. From our evaluation we found that although the addressed problems are extremely challenging, the proposed system is able to rank the winner in the top 3 best predicted scores for 5 out of 10 Miss Universe competitions.

Automatic video keyword generation is one of the key ingredients in reducing the burden of security officers in analyzing surveillance videos. Keywords or attributes are generally chosen manually based on expert knowledge of surveillance. Most existing works primarily aim at either supervised learning approaches relying on extensive manual labelling or hierarchical probabilistic models that assume the features are extracted using the bag-of-words approach; thus limiting the utilization of the other features. To address this, we turn our attention to automatic attribute discovery approaches. However, it is not clear which automatic discovery approach can discover the most meaningful attributes. Furthermore, little research has been done on how to compare and choose the best automatic attribute discovery methods. In this paper, we propose a novel approach, based on the shared structure exhibited amongst meaningful attributes, that enables us to compare between different automatic attribute discovery approaches.We then validate our approach by comparing various attribute discovery methods such as PiCoDeS on two attribute datasets. The evaluation shows that our approach is able to select the automatic discovery approach that discovers the most meaningful attributes. We then employ the best discovery approach to generate keywords for videos recorded from a surveillance system. This work shows it is possible to massively reduce the amount of manual work in generating video keywords without limiting ourselves to a particular video feature descriptor., Comment: 7 pages, ISBA 2016. arXiv admin note: text overlap with arXiv:1602.01940

Many automatic attribute discovery methods have been developed to extract a set of visual attributes from images for various tasks. However, despite good performance in some image classification tasks, it is difficult to evaluate whether these methods discover meaningful attributes and which one is the best to find the attributes for image descriptions. An intuitive way to evaluate this is to manually verify whether consistent identifiable visual concepts exist to distinguish between positive and negative images of an attribute. This manual checking is tedious, labor intensive and expensive and it is very hard to get quantitative comparisons between different methods. In this work, we tackle this problem by proposing an attribute meaningfulness metric, that can perform automatic evaluation on the meaningfulness of attribute sets as well as achieving quantitative comparisons. We apply our proposed metric to recent automatic attribute discovery methods and popular hashing methods on three attribute datasets. A user study is also conducted to validate the effectiveness of the metric. In our evaluation, we gleaned some insights that could be beneficial in developing automatic attribute discovery methods to generate meaningful attributes. To the best of our knowledge, this is the first work to quantitatively measure the semantic content of automatically discovered attributes., Comment: 9 pages, WACV 2016

We present a comparative evaluation of various techniques for action recognition while keeping as many variables as possible controlled. We employ two categories of Riemannian manifolds: symmetric positive definite matrices and linear subspaces. For both categories we use their corresponding nearest neighbour classifiers, kernels, and recent kernelised sparse representations. We compare against traditional action recognition techniques based on Gaussian mixture models and Fisher vectors (FVs). We evaluate these action recognition techniques under ideal conditions, as well as their sensitivity in more challenging conditions (variations in scale and translation). Despite recent advancements for handling manifolds, manifold based techniques obtain the lowest performance and their kernel representations are more unstable in the presence of challenging conditions. The FV approach obtains the highest accuracy under ideal conditions. Moreover, FV best deals with moderate scale and translation changes.

Reformulating computer vision problems over Riemannian manifolds has demonstrated superior performance in various computer vision applications. This is because visual data often forms a special structure lying on a lower dimensional space embedded in a higher dimensional space. However, since these manifolds belong to non-Euclidean topological spaces, exploiting their structures is computationally expensive, especially when one considers the clustering analysis of massive amounts of data. To this end, we propose an efficient framework to address the clustering problem on Riemannian manifolds. This framework implements random projections for manifold points via kernel space, which can preserve the geometric structure of the original space, but is computationally efficient. Here, we introduce three methods that follow our framework. We then validate our framework on several computer vision applications by comparing against popular clustering methods on Riemannian manifolds. Experimental results demonstrate that our framework maintains the performance of the clustering whilst massively reducing computational complexity by over two orders of magnitude in some cases.

Recently, there has been a growing interest in developing Computer Aided Diagnostic (CAD) systems for improving the reliability and consistency of pathology test results. This paper describes a novel CAD system for the Anti-Nuclear Antibody (ANA) test via Indirect Immunofluorescence protocol on Human Epithelial Type 2 (HEp-2) cells. While prior works have primarily focused on classifying cell images extracted from ANA specimen images, this work takes a further step by focussing on the specimen image classification problem itself. Our system is able to efficiently classify specimen images as well as producing meaningful descriptions of ANA pattern class which helps physicians to understand the differences between various ANA patterns. We achieve this goal by designing a specimen-level image descriptor that: (1) is highly discriminative; (2) has small descriptor length and (3) is semantically meaningful at the cell level. In our work, a specimen image descriptor is represented by its overall cell attribute descriptors. As such, we propose two max-margin based learning schemes to discover cell attributes whilst still maintaining the discrimination of the specimen image descriptor. Our learning schemes differ from the existing discriminative attribute learning approaches as they primarily focus on discovering image-level attributes. Comparative evaluations were undertaken to contrast the proposed approach to various state-of-the-art approaches on a novel HEp-2 cell dataset which was specifically proposed for the specimen-level classification. Finally, we showcase the ability of the proposed approach to provide textual descriptions to explain ANA patterns., Comment: WACV 2014: IEEE Winter Conference on Applications of Computer Vision

This paper describes a novel system for automatic classification of images obtained from Anti-Nuclear Antibody (ANA) pathology tests on Human Epithelial type 2 (HEp-2) cells using the Indirect Immunofluorescence (IIF) protocol. The IIF protocol on HEp-2 cells has been the hallmark method to identify the presence of ANAs, due to its high sensitivity and the large range of antigens that can be detected. However, it suffers from numerous shortcomings, such as being subjective as well as time and labour intensive. Computer Aided Diagnostic (CAD) systems have been developed to address these problems, which automatically classify a HEp-2 cell image into one of its known patterns (eg. speckled, homogeneous). Most of the existing CAD systems use handpicked features to represent a HEp-2 cell image, which may only work in limited scenarios. We propose a novel automatic cell image classification method termed Cell Pyramid Matching (CPM), which is comprised of regional histograms of visual words coupled with the Multiple Kernel Learning framework. We present a study of several variations of generating histograms and show the efficacy of the system on two publicly available datasets: the ICPR HEp-2 cell classification contest dataset and the SNPHEp-2 dataset., Comment: arXiv admin note: substantial text overlap with arXiv:1304.1262

Recent advances suggest that encoding images through Symmetric Positive Definite (SPD) matrices and then interpreting such matrices as points on Riemannian manifolds can lead to increased classification performance. Taking into account manifold geometry is typically done via (1) embedding the manifolds in tangent spaces, or (2) embedding into Reproducing Kernel Hilbert Spaces (RKHS). While embedding into tangent spaces allows the use of existing Euclidean-based learning algorithms, manifold shape is only approximated which can cause loss of discriminatory information. The RKHS approach retains more of the manifold structure, but may require non-trivial effort to kernelise Euclidean-based learning algorithms. In contrast to the above approaches, in this paper we offer a novel solution that allows SPD matrices to be used with unmodified Euclidean-based learning algorithms, with the true manifold shape well-preserved. Specifically, we propose to project SPD matrices using a set of random projection hyperplanes over RKHS into a random projection space, which leads to representing each matrix as a vector of projection coefficients. Experiments on face recognition, person re-identification and texture classification show that the proposed approach outperforms several recent methods, such as Tensor Sparse Coding, Histogram Plus Epitome, Riemannian Locality Preserving Projection and Relational Divergence Classification., Comment: IEEE Winter Conference on Applications of Computer Vision (WACV), 2014

Traditional nearest points methods use all the samples in an image set to construct a single convex or affine hull model for classification. However, strong artificial features and noisy data may be generated from combinations of training samples when significant intra-class variations and/or noise occur in the image set. Existing multi-model approaches extract local models by clustering each image set individually only once, with fixed clusters used for matching with various image sets. This may not be optimal for discrimination, as undesirable environmental conditions (eg. illumination and pose variations) may result in the two closest clusters representing different characteristics of an object (eg. frontal face being compared to non-frontal face). To address the above problem, we propose a novel approach to enhance nearest points based methods by integrating affine/convex hull classification with an adapted multi-model approach. We first extract multiple local convex hulls from a query image set via maximum margin clustering to diminish the artificial variations and constrain the noise in local convex hulls. We then propose adaptive reference clustering (ARC) to constrain the clustering of each gallery image set by forcing the clusters to have resemblance to the clusters in the query image set. By applying ARC, noisy clusters in the query set can be discarded. Experiments on Honda, MoBo and ETH-80 datasets show that the proposed method outperforms single model approaches and other recent techniques, such as Sparse Approximated Nearest Points, Mutual Subspace Method and Manifold Discriminant Analysis., Comment: IEEE Winter Conference on Applications of Computer Vision (WACV), 2014

The Anti-Nuclear Antibody (ANA) clinical pathology test is commonly used to identify the existence of various diseases. A hallmark method for identifying the presence of ANAs is the Indirect Immunofluorescence method on Human Epithelial (HEp-2) cells, due to its high sensitivity and the large range of antigens that can be detected. However, the method suffers from numerous shortcomings, such as being subjective as well as time and labour intensive. Computer Aided Diagnostic (CAD) systems have been developed to address these problems, which automatically classify a HEp-2 cell image into one of its known patterns (eg., speckled, homogeneous). Most of the existing CAD systems use handpicked features to represent a HEp-2 cell image, which may only work in limited scenarios. In this paper, we propose a cell classification system comprised of a dual-region codebook-based descriptor, combined with the Nearest Convex Hull Classifier. We evaluate the performance of several variants of the descriptor on two publicly available datasets: ICPR HEp-2 cell classification contest dataset and the new SNPHEp-2 dataset. To our knowledge, this is the first time codebook-based descriptors are applied and studied in this domain. Experiments show that the proposed system has consistent high performance and is more robust than two recent CAD systems.