Energy-based latent variable models (EBLVMs) are more expressive than conventional energy-based models. However, its potential on visual tasks are limited by its training process based on maximum likelihood estimate that requires sampling from two intractable distributions. In this paper, we propose Bi-level doubly variational learning (BiDVL), which is based on a new bi-level optimization framework and two tractable variational distributions to facilitate learning EBLVMs. Particularly, we lead a decoupled EBLVM consisting of a marginal energy-based distribution and a structural posterior to handle the difficulties when learning deep EBLVMs on images. By choosing a symmetric KL divergence in the lower level of our framework, a compact BiDVL for visual tasks can be obtained. Our model achieves impressive image generation performance over related works. It also demonstrates the significant capacity of testing image reconstruction and out-of-distribution detection.
Music can be represented in multiple forms, such as in the audio form as a recording of a performance, in the symbolic form as a computer readable score, or in the image form as a scan of the sheet music. Music synchronisation provides a way to navigate among multiple representations of music in a unified manner by generating an accurate mapping between them, lending itself applicable to a myriad of domains like music education, performance analysis, automatic accompaniment and music editing. Traditional synchronisation methods compute alignment using knowledge-driven and stochastic approaches, typically employing handcrafted features. These methods are often unable to generalise well to different instruments, acoustic environments and recording conditions, and normally assume complete structural agreement between the performances and the scores. This PhD furthers the development of performance-score synchronisation research by proposing data-driven, context-aware alignment approaches, on three fronts: Firstly, I replace the handcrafted features by employing a metric learning based approach that is adaptable to different acoustic settings and performs well in data-scarce conditions. Secondly, I address the handling of structural differences between the performances and scores, which is a common limitation of standard alignment methods. Finally, I eschew the reliance on both feature engineering and dynamic programming, and propose a completely data-driven synchronisation method that computes alignments using a neural framework, whilst also being robust to structural differences between the performances and scores.
Cross-view image matches have been widely explored on terrestrial image localization using aerial images from drones or satellites. This study expands the cross-view image match idea and proposes a cross-domain and cross-view localization framework. The method identifies the correlation between color aerial images and underwater acoustic images to improve the localization of underwater vehicles that travel in partially structured environments such as harbors and marinas. The approach is validated on a real dataset acquired by an underwater vehicle in a marina. The results show an improvement in the localization when compared to the dead reckoning of the vehicle.
In this paper, we aim to improve the performance of in-the-wild Facial Expression Recognition (FER) by exploiting semi-supervised learning. Large-scale labeled data and deep learning methods have greatly improved the performance of image recognition. However, the performance of FER is still not ideal due to the lack of training data and incorrect annotations (e.g., label noises). Among existing in-the-wild FER datasets, reliable ones contain insufficient data to train robust deep models while large-scale ones are annotated in lower quality. To address this problem, we propose a semi-supervised learning algorithm named Progressive Teacher (PT) to utilize reliable FER datasets as well as large-scale unlabeled expression images for effective training. On the one hand, PT introduces semi-supervised learning method to relieve the shortage of data in FER. On the other hand, it selects useful labeled training samples automatically and progressively to alleviate label noise. PT uses selected clean labeled data for computing the supervised classification loss and unlabeled data for unsupervised consistency loss. Experiments on widely-used databases RAF-DB and FERPlus validate the effectiveness of our method, which achieves state-of-the-art performance with accuracy of 89.57% on RAF-DB. Additionally, when the synthetic noise rate reaches even 30%, the performance of our PT algorithm only degrades by 4.37%.
Conventional image reconstruction models for lensless cameras often assume that each measurement results from convolving a given scene with a single experimentally measured point-spread function. These image reconstruction models fall short in simulating lensless cameras truthfully as these models are not sophisticated enough to account for optical aberrations or scenes with depth variations. Our work shows that learning a supervised primal-dual reconstruction method results in image quality matching state of the art in the literature without demanding a large network capacity. This improvement stems from our primary finding that embedding learnable forward and adjoint models in a learned primal-dual optimization framework can even improve the quality of reconstructed images (+5dB PSNR) compared to works that do not correct for the model error. In addition, we built a proof-of-concept lensless camera prototype that uses a pseudo-random phase mask to demonstrate our point. Finally, we share the extensive evaluation of our learned model based on an open dataset and a dataset from our proof-of-concept lensless camera prototype.
An omni-directional image (ODI) is the image that has a field of view covering the entire sphere around the camera. The ODIs have begun to be used in a wide range of fields such as virtual reality (VR), robotics, and social network services. Although the contents using ODI have increased, the available images and videos are still limited, compared with widespread snapshot images. A large number of ODIs are desired not only for the VR contents, but also for training deep learning models for ODI. For these purposes, a novel computer vision task to generate ODI from a single snapshot image is proposed in this paper. To tackle this problem, the conditional generative adversarial network was applied in combination with class-conditioned convolution layers. With this novel task, VR images and videos will be easily created even with a smartphone camera.
The success of deep neural networks in image classification and learning can be partly attributed to the features they extract from images. It is often speculated about the properties of a low-dimensional manifold that models extract and learn from images. However, there is not sufficient understanding about this low-dimensional space based on theory or empirical evidence. For image classification models, their last hidden layer is the one where images of each class is separated from other classes and it also has the least number of features. Here, we develop methods and formulations to study that feature space for any model. We study the partitioning of the domain in feature space, identify regions guaranteed to have certain classifications, and investigate its implications for the pixel space. We observe that geometric arrangements of decision boundaries in feature space is significantly different compared to pixel space, providing insights about adversarial vulnerabilities, image morphing, extrapolation, ambiguity in classification, and the mathematical understanding of image classification models.
Developing technology and changing lifestyles have made online grocery delivery applications an indispensable part of urban life. Since the beginning of the COVID-19 pandemic, the demand for such applications has dramatically increased, creating new competitors that disrupt the market. An increasing level of competition might prompt companies to frequently restructure their marketing and product pricing strategies. Therefore, identifying the change patterns in product prices and sales volumes would provide a competitive advantage for the companies in the marketplace. In this paper, we investigate alternative clustering methodologies to group the products based on the price patterns and sales volumes. We propose a novel distance metric that takes into account how product prices and sales move together rather than calculating the distance using numerical values. We compare our approach with traditional clustering algorithms, which typically rely on generic distance metrics such as Euclidean distance, and image clustering approaches that aim to group data by capturing its visual patterns. We evaluate the performances of different clustering algorithms using our custom evaluation metric as well as Calinski Harabasz and Davies Bouldin indices, which are commonly used internal validity metrics. We conduct our numerical study using a propriety price dataset from an online food and grocery delivery company, and the publicly available Favorita sales dataset. We find that our proposed clustering approach and image clustering both perform well for finding the products with similar price and sales patterns within large datasets.
Interactive garment retrieval (IGR) aims to retrieve a target garment image based on a reference garment image along with user feedback on what to change on the reference garment. Two IGR tasks have been studied extensively: text-guided garment retrieval (TGR) and visually compatible garment retrieval (VCR). The user feedback for the former indicates what semantic attributes to change with the garment category preserved, while the category is the only thing to be changed explicitly for the latter, with an implicit requirement on style preservation. Despite the similarity between these two tasks and the practical need for an efficient system tackling both, they have never been unified and modeled jointly. In this paper, we propose a Unified Interactive Garment Retrieval (UIGR) framework to unify TGR and VCR. To this end, we first contribute a large-scale benchmark suited for both problems. We further propose a strong baseline architecture to integrate TGR and VCR in one model. Extensive experiments suggest that unifying two tasks in one framework is not only more efficient by requiring a single model only, it also leads to better performance. Code and datasets are available at https://github.com/BrandonHanx/CompFashion.
Camera relocalization is the key component of simultaneous localization and mapping (SLAM) systems. This paper proposes a learning-based approach, named Sparse Spatial Scene Embedding with Graph Neural Networks (S3E-GNN), as an end-to-end framework for efficient and robust camera relocalization. S3E-GNN consists of two modules. In the encoding module, a trained S3E network encodes RGB images into embedding codes to implicitly represent spatial and semantic embedding code. With embedding codes and the associated poses obtained from a SLAM system, each image is represented as a graph node in a pose graph. In the GNN query module, the pose graph is transformed to form a embedding-aggregated reference graph for camera relocalization. We collect various scene datasets in the challenging environments to perform experiments. Our results demonstrate that S3E-GNN method outperforms the traditional Bag-of-words (BoW) for camera relocalization due to learning-based embedding and GNN powered scene matching mechanism.