In computer vision, finding point correspondence among images plays an important role in many applications, such as image stitching, image retrieval, visual localization, etc. Most of the research worksfocus on the matching of local feature before a sampling method is employed, such as RANSAC, to verify initial matching results via repeated fitting of certain global transformation among the images. However, incorrect matches may still exist, while careful examination of such problems is often skipped. Accordingly, a geometrically constrained algorithm is proposed in this work to verify the correctness of initially matched SIFT keypoints based on view-invariant cross-ratios (CRs). By randomly forming pentagons from these keypoints and matching their shape and location among images with CRs, robust planar region estimation can be achieved efficiently for the above verification, while correct and incorrect matches of keypoints can be examined easily with respect to those shape and location matched pentagons. Experimental results show that satisfactory results can be obtained for various scenes with single as well as multiple planar regions.
Much computer vision research has focused on natural images, but technical documents typically consist of abstract images, such as charts, drawings, diagrams, and schematics. How well do general web search engines discover abstract images? Recent advancements in computer vision and machine learning have led to the rise of reverse image search engines. Where conventional search engines accept a text query and return a set of document results, including images, a reverse image search accepts an image as a query and returns a set of images as results. This paper evaluates how well common reverse image search engines discover abstract images. We conducted an experiment leveraging images from Wikimedia Commons, a website known to be well indexed by Baidu, Bing, Google, and Yandex. We measure how difficult an image is to find again (retrievability), what percentage of images returned are relevant (precision), and the average number of results a visitor must review before finding the submitted image (mean reciprocal rank). When trying to discover the same image again among similar images, Yandex performs best. When searching for pages containing a specific image, Google and Yandex outperform the others when discovering photographs with precision scores ranging from 0.8191 to 0.8297, respectively. In both of these cases, Google and Yandex perform better with natural images than with abstract ones achieving a difference in retrievability as high as 54\% between images in these categories. These results affect anyone applying common web search engines to search for technical documents that use abstract images.
This work investigates the use of large-scale, English-only pre-trained models (CLIP and HuBERT) for multilingual image-speech retrieval. For non-English image-speech retrieval, we outperform the current state-of-the-art performance by a wide margin both when training separate models for each language, and with a single model which processes speech in all three languages. We identify key differences in model behavior and performance between English and non-English settings, attributable to the English-only pre-training of CLIP and HuBERT, and investigate how fine-tuning the pre-trained models impacts these differences. Finally, we show that our models can be used for mono- and cross-lingual speech-text retrieval and cross-lingual speech-speech retrieval, despite never having seen any parallel speech-text or speech-speech data during training.
Recent visuolinguistic pre-trained models show promising progress on various end tasks such as image retrieval and video captioning. Yet, they fail miserably on the recently proposed Winoground dataset, which challenges models to match paired images and English captions, with items constructed to overlap lexically but differ in meaning (e.g., "there is a mug in some grass" vs. "there is some grass in a mug"). By annotating the dataset using new fine-grained tags, we show that solving the Winoground task requires not just compositional language understanding, but a host of other abilities like commonsense reasoning or locating small, out-of-focus objects in low-resolution images. In this paper, we identify the dataset's main challenges through a suite of experiments on related tasks (probing task, image retrieval task), data augmentation, and manual inspection of the dataset. Our analysis suggests that a main challenge in visuolinguistic models may lie in fusing visual and textual representations, rather than in compositional language understanding. We release our annotation and code at https://github.com/ajd12342/why-winoground-hard .
The tremendous success of CLIP (Radford et al., 2021) has promoted the research and application of contrastive learning for vision-language pretraining. In this work, we construct a large-scale dataset of image-text pairs in Chinese, where most data are retrieved from publicly available datasets, and we pretrain Chinese CLIP models on the new dataset. We develop 5 Chinese CLIP models of multiple sizes, spanning from 77 to 958 million parameters. Furthermore, we propose a two-stage pretraining method, where the model is first trained with the image encoder frozen and then trained with all parameters being optimized, to achieve enhanced model performance. Our comprehensive experiments demonstrate that Chinese CLIP can achieve the state-of-the-art performance on MUGE, Flickr30K-CN, and COCO-CN in the setups of zero-shot learning and finetuning, and it is able to achieve competitive performance in zero-shot image classification based on the evaluation on the ELEVATER benchmark (Li et al., 2022). We have released our codes, models, and demos in https://github.com/OFA-Sys/Chinese-CLIP
Several techniques to map various types of components, such as words, attributes, and images, into the embedded space have been studied. Most of them estimate the embedded representation of target entity as a point in the projective space. Some models, such as Word2Gauss, assume a probability distribution behind the embedded representation, which enables the spread or variance of the meaning of embedded target components to be captured and considered in more detail. We examine the method of estimating embedded representations as probability distributions for the interpretation of fashion-specific abstract and difficult-to-understand terms. Terms, such as "casual," "adult-casual,'' "beauty-casual," and "formal," are extremely subjective and abstract and are difficult for both experts and non-experts to understand, which discourages users from trying new fashion. We propose an end-to-end model called dual Gaussian visual-semantic embedding, which maps images and attributes in the same projective space and enables the interpretation of the meaning of these terms by its broad applications. We demonstrate the effectiveness of the proposed method through multifaceted experiments involving image and attribute mapping, image retrieval and re-ordering techniques, and a detailed theoretical/analytical discussion of the distance measure included in the loss function.
Sketches have been used to conceptualise and depict visual objects from pre-historic times. Sketch research has flourished in the past decade, particularly with the proliferation of touchscreen devices. Much of the utilisation of sketch has been anchored around the fact that it can be used to delineate visual concepts universally irrespective of age, race, language, or demography. The fine-grained interactive nature of sketches facilitates the application of sketches to various visual understanding tasks, like image retrieval, image-generation or editing, segmentation, 3D-shape modelling etc. However, sketches are highly abstract and subjective based on the perception of individuals. Although most agree that sketches provide fine-grained control to the user to depict a visual object, many consider sketching a tedious process due to their limited sketching skills compared to other query/support modalities like text/tags. Furthermore, collecting fine-grained sketch-photo association is a significant bottleneck to commercialising sketch applications. Therefore, this thesis aims to progress sketch-based visual understanding towards more practicality.
People post their opinions and experiences on social media, yielding rich databases of end-users' sentiments. This paper shows to what extent machine learning can analyze and structure these databases. An automated data analysis pipeline is deployed to provide insights into user-generated content for researchers in other domains. First, the domain expert can select an image and a term of interest. Then, the pipeline uses image retrieval to find all images showing similar content and applies aspect-based sentiment analysis to outline users' opinions about the selected term. As part of an interdisciplinary project between architecture and computer science researchers, an empirical study of Hamburg's Elbphilharmonie was conveyed. Therefore, we selected 300 thousand posts with the hashtag \enquote{\texttt{hamburg}} from the platform Flickr. Image retrieval methods generated a subset of slightly more than 1.5 thousand images displaying the Elbphilharmonie. We found that these posts mainly convey a neutral or positive sentiment towards it. With this pipeline, we suggest a new semantic computing method that offers novel insights into end-users opinions, e.g., for architecture domain experts.
With the increasing availability of large digitized fine art collections, automated analysis and classification of paintings is becoming an interesting area of research. However, due to domain specificity, implicit subjectivity, and pervasive nuances that vaguely separate art movements, analyzing art using machine learning techniques poses significant challenges. Residual networks, or variants thereof, are one the most popular tools for image classification tasks, which can extract relevant features for well-defined classes. In this case study, we focus on the classification of a selected painting 'Portrait of the Painter Charles Bruni' by Johann Kupetzky and the analysis of the performance of the proposed classifier. We show that the features extracted during residual network training can be useful for image retrieval within search systems in online art collections.
Instance-level image retrieval in fashion is a challenging issue owing to its increasing importance in real-scenario visual fashion search. Cross-domain fashion retrieval aims to match the unconstrained customer images as queries for photographs provided by retailers; however, it is a difficult task due to a wide range of consumer-to-shop (C2S) domain discrepancies and also considering that clothing image is vulnerable to various non-rigid deformations. To this end, we propose a novel multi-scale and multi-granularity feature learning network (MMFL-Net), which can jointly learn global-local aggregation feature representations of clothing images in a unified framework, aiming to train a cross-domain model for C2S fashion visual similarity. First, a new semantic-spatial feature fusion part is designed to bridge the semantic-spatial gap by applying top-down and bottom-up bidirectional multi-scale feature fusion. Next, a multi-branch deep network architecture is introduced to capture global salient, part-informed, and local detailed information, and extracting robust and discrimination feature embedding by integrating the similarity learning of coarse-to-fine embedding with the multiple granularities. Finally, the improved trihard loss, center loss, and multi-task classification loss are adopted for our MMFL-Net, which can jointly optimize intra-class and inter-class distance and thus explicitly improve intra-class compactness and inter-class discriminability between its visual representations for feature learning. Furthermore, our proposed model also combines the multi-task attribute recognition and classification module with multi-label semantic attributes and product ID labels. Experimental results demonstrate that our proposed MMFL-Net achieves significant improvement over the state-of-the-art methods on the two datasets, DeepFashion-C2S and Street2Shop.
Multimodal tasks in the fashion domain have significant potential for e-commerce, but involve challenging vision-and-language learning problems - e.g., retrieving a fashion item given a reference image plus text feedback from a user. Prior works on multimodal fashion tasks have either been limited by the data in individual benchmarks, or have leveraged generic vision-and-language pre-training but have not taken advantage of the characteristics of fashion data. Additionally, these works have mainly been restricted to multimodal understanding tasks. To address these gaps, we make two key contributions. First, we propose a novel fashion-specific pre-training framework based on weakly-supervised triplets constructed from fashion image-text pairs. We show the triplet-based tasks are an effective addition to standard multimodal pre-training tasks. Second, we propose a flexible decoder-based model architecture capable of both fashion retrieval and captioning tasks. Together, our model design and pre-training approach are competitive on a diverse set of fashion tasks, including cross-modal retrieval, image retrieval with text feedback, image captioning, relative image captioning, and multimodal categorization.
The Vision-Language Pre-training (VLP) models like CLIP have gained popularity in recent years. However, many works found that the social biases hidden in CLIP easily manifest in downstream tasks, especially in image retrieval, which can have harmful effects on human society. In this work, we propose FairCLIP to eliminate the social bias in CLIP-based image retrieval without damaging the retrieval performance achieving the compatibility between the debiasing effect and the retrieval performance. FairCLIP is divided into two steps: Attribute Prototype Learning (APL) and Representation Neutralization (RN). In the first step, we extract the concepts needed for debiasing in CLIP. We use the query with learnable word vector prefixes as the extraction structure. In the second step, we first divide the attributes into target and bias attributes. By analysis, we find that both attributes have an impact on the bias. Therefore, we try to eliminate the bias by using Re-Representation Matrix (RRM) to achieve the neutralization of the representation. We compare the debiasing effect and retrieval performance with other methods, and experiments demonstrate that FairCLIP can achieve the best compatibility. Although FairCLIP is used to eliminate bias in image retrieval, it achieves the neutralization of the representation which is common to all CLIP downstream tasks. This means that FairCLIP can be applied as a general debiasing method for other fairness issues related to CLIP.
Weakly-supervised vision-language (V-L) pre-training (W-VLP) aims at learning cross-modal alignment with little or no paired data, such as aligned images and captions. Recent W-VLP methods, which pair visual features with object tags, help achieve performances comparable with some VLP models trained with aligned pairs in various V-L downstream tasks. This, however, is not the case in cross-modal retrieval (XMR). We argue that the learning of such a W-VLP model is curbed and biased by the object tags of limited semantics. We address the lack of paired V-L data for model supervision with a novel Visual Vocabulary based Feature Hallucinator (WFH), which is trained via weak supervision as a W-VLP model, not requiring images paired with captions. WFH generates visual hallucinations from texts, which are then paired with the originally unpaired texts, allowing more diverse interactions across modalities. Empirically, WFH consistently boosts the prior W-VLP works, e.g. U-VisualBERT (U-VB), over a variety of V-L tasks, i.e. XMR, Visual Question Answering, etc. Notably, benchmarked with recall@{1,5,10}, it consistently improves U-VB on image-to-text and text-to-image retrieval on two popular datasets Flickr30K and MSCOCO. Meanwhile, it gains by at least 14.5% in cross-dataset generalization tests on these XMR tasks. Moreover, in other V-L downstream tasks considered, our WFH models are on par with models trained with paired V-L data, revealing the utility of unpaired data. These results demonstrate greater generalization of the proposed W-VLP model with WFH.
This paper introduces a structured, adaptive-length deep representation called Neural Eigenmap. Unlike prior spectral methods such as Laplacian Eigenmap that operate in a nonparametric manner, Neural Eigenmap leverages NeuralEF to parametrically model eigenfunctions using a neural network. We show that, when the eigenfunction is derived from positive relations in a data augmentation setup, applying NeuralEF results in an objective function that resembles those of popular self-supervised learning methods, with an additional symmetry-breaking property that leads to \emph{structured} representations where features are ordered by importance. We demonstrate using such representations as adaptive-length codes in image retrieval systems. By truncation according to feature importance, our method requires up to $16\times$ shorter representation length than leading self-supervised learning ones to achieve similar retrieval performance. We further apply our method to graph data and report strong results on a node representation learning benchmark with more than one million nodes.
Current person image retrieval methods have achieved great improvements in accuracy metrics. However, they rarely describe the reliability of the prediction. In this paper, we propose an Uncertainty-Aware Learning (UAL) method to remedy this issue. UAL aims at providing reliability-aware predictions by considering data uncertainty and model uncertainty simultaneously. Data uncertainty captures the ``noise" inherent in the sample, while model uncertainty depicts the model's confidence in the sample's prediction. Specifically, in UAL, (1) we propose a sampling-free data uncertainty learning method to adaptively assign weights to different samples during training, down-weighting the low-quality ambiguous samples. (2) we leverage the Bayesian framework to model the model uncertainty by assuming the parameters of the network follow a Bernoulli distribution. (3) the data uncertainty and the model uncertainty are jointly learned in a unified network, and they serve as two fundamental criteria for the reliability assessment: if a probe is high-quality (low data uncertainty) and the model is confident in the prediction of the probe (low model uncertainty), the final ranking will be assessed as reliable. Experiments under the risk-controlled settings and the multi-query settings show the proposed reliability assessment is effective. Our method also shows superior performance on three challenging benchmarks under the vanilla single query settings.
Vision transformers have achieved remarkable progress in vision tasks such as image classification and detection. However, in instance-level image retrieval, transformers have not yet shown good performance compared to convolutional networks. We propose a number of improvements that make transformers outperform the state of the art for the first time. (1) We show that a hybrid architecture is more effective than plain transformers, by a large margin. (2) We introduce two branches collecting global (classification token) and local (patch tokens) information, from which we form a global image representation. (3) In each branch, we collect multi-layer features from the transformer encoder, corresponding to skip connections across distant layers. (4) We enhance locality of interactions at the deeper layers of the encoder, which is the relative weakness of vision transformers. We train our model on all commonly used training sets and, for the first time, we make fair comparisons separately per training set. In all cases, we outperform previous models based on global representation. Public code is available at https://github.com/dealicious-inc/DToP.
Scene graphs provide structured semantic understanding beyond images. For downstream tasks, such as image retrieval, visual question answering, visual relationship detection, and even autonomous vehicle technology, scene graphs can not only distil complex image information but also correct the bias of visual models using semantic-level relations, which has broad application prospects. However, the heavy labour cost of constructing graph annotations may hinder the application of PSG in practical scenarios. Inspired by the observation that people usually identify the subject and object first and then determine the relationship between them, we proposed to decouple the scene graphs generation task into two sub-tasks: 1) an image segmentation task to pick up the qualified objects. 2) a restricted auto-regressive text generation task to generate the relation between given objects. Therefore, in this work, we introduce image semantic relation generation (ISRG), a simple but effective image-to-text model, which achieved 31 points on the OpenPSG dataset and outperforms strong baselines respectively by 16 points (ResNet-50) and 5 points (CLIP).
The Google Universal Image Embedding (GUIE) Challenge is one of the first competitions in multi-domain image representations in the wild, covering a wide distribution of objects: landmarks, artwork, food, etc. This is a fundamental computer vision problem with notable applications in image retrieval, search engines and e-commerce. In this work, we explain our 4th place solution to the GUIE Challenge, and our "bag of tricks" to fine-tune zero-shot Vision Transformers (ViT) pre-trained using CLIP.
Representation learning for sketch-based image retrieval has mostly been tackled by learning embeddings that discard modality-specific information. As instances from different modalities can often provide complementary information describing the underlying concept, we propose a cross-attention framework for Vision Transformers (XModalViT) that fuses modality-specific information instead of discarding them. Our framework first maps paired datapoints from the individual photo and sketch modalities to fused representations that unify information from both modalities. We then decouple the input space of the aforementioned modality fusion network into independent encoders of the individual modalities via contrastive and relational cross-modal knowledge distillation. Such encoders can then be applied to downstream tasks like cross-modal retrieval. We demonstrate the expressive capacity of the learned representations by performing a wide range of experiments and achieving state-of-the-art results on three fine-grained sketch-based image retrieval benchmarks: Shoe-V2, Chair-V2 and Sketchy. Implementation is available at https://github.com/abhrac/xmodal-vit.
This paper aims to investigate representation learning for large scale visual place recognition, which consists of determining the location depicted in a query image by referring to a database of reference images. This is a challenging task due to the large-scale environmental changes that can occur over time (i.e., weather, illumination, season, traffic, occlusion). Progress is currently challenged by the lack of large databases with accurate ground truth. To address this challenge, we introduce GSV-Cities, a new image dataset providing the widest geographic coverage to date with highly accurate ground truth, covering more than 40 cities across all continents over a 14-year period. We subsequently explore the full potential of recent advances in deep metric learning to train networks specifically for place recognition, and evaluate how different loss functions influence performance. In addition, we show that performance of existing methods substantially improves when trained on GSV-Cities. Finally, we introduce a new fully convolutional aggregation layer that outperforms existing techniques, including GeM, NetVLAD and CosPlace, and establish a new state-of-the-art on large-scale benchmarks, such as Pittsburgh, Mapillary-SLS, SPED and Nordland. The dataset and code are available for research purposes at https://github.com/amaralibey/gsv-cities.