Local Semantic Structure Captured and Instance Discriminated by Unsupervised Hashing

The classification of marine plankton images is of great significance in ecological studies and environmental monitoring. In practical applications, plankton image classification faces several challenges, including sample imbalance, distinguishing between-class and within-class differences, and recognizing fine-grained features. To address these issues, we propose a few-shot self-supervised transfer learning (FSTL) framework. In FSTL, we design a new loss function that incorporates both supervised and self-supervised learning. The core of FSTL is a hybrid learning objective that integrates self-supervised contrastive learning for robust feature representation. Operating within a transfer learning paradigm, FSTL effectively adapts knowledge from head-classes to boost the few-shot classification performance on tail-classes. We applied FSTL to two datasets, in which plankton images were collected from Daya Bay and provided by the Woods Hole Oceanographic Institution (WHOI) datasets respectively. The experimental results demonstrated that our method showed better adaptability in the classification of plankton images. The findings of this study not only apply to the classification of plankton images but also offer the potential for classifying small-sample categories within long-tailed datasets.

Hashing technology, one of the most efficient approximate nearest neighbor searching methods due to its fast query speed and low storage cost, has been widely used in image retrieval. Recently, unsupervised deep hashing methods have attracted more and more attention due to the lack of labels in real applications. Most unsupervised hashing methods usually construct a similarity matrix with the features extracted from the images, and then guide the hash code learning with this similarity matrix. However, in unsupervised scenario, such similarity matrix may be unreliable due to the affect of noise and irrelevant objects in images. In this paper, we propose a novel unsupervised deep hashing method called Deep Structured Hashing (DSH). In the new method, we first learn both continuous and binary structured similarity matrices with explicit cluster structure to better preserve the semantic structure, where the binary one preserves the coarse-grained semantic structure while the continuous one preserves the fine-grained semantic structure. And then jointly optimize three kinds of losses to learn high quality hash codes. Extensive experiments on three benchmark datasets show the superior retrieval performance of our proposed method.

The booming demands for cross-modal retrieval tasks can often bring in its wake the development of retrieval technologies, as people turn to pursuing a more effective way to improve the performance of search results in both accuracy and efficiency, for example by unsupervised cross-modal hashing. It’s worth noting that most of the cross-modal hashing methods focus on utilizing merely one approach to generate hash codes. However, each approach has its own intrinsic drawback, which would inevitably diminish the quality of hash codes. In this paper, we propose a state-of-the-art model named Soft Constraint Hashing (SCH), using a special soft constraint term defined as an “information tunnel” to achieve the goal that conveys information from one approach to another. In particular, this “tunnel” can eliminate potential data noises to some extent and bridge the gap between two unsupervised discrete hashing allocation approaches to simultaneously reinforce the quality of hash codes. The empirical results on publicly available datasets illustrate that our proposed model outperforms all the existing unsupervised cross-model hashing methods.

Due to their low storage capacity and fast retrieval speed, hashing techniques have received much attention in cross-modal retrieval. However, there are some issues that need to be further explored. First, some existing hashing methods use the labels to construct the semantic similarity matrix between pairwise data, ignoring the potential manifold structure between heterogeneous data. Second, some existing methods underestimate the importance of multi-label and the gaps between different class labels, making the learned hash codes less discriminative. Third, few of them embed both manifold and balanced structures within the same model, and the relaxation of discrete constraints will lead to an increasing quantization error. To mitigate these problems, this paper proposes a novel supervised hashing method, termed Label Guided Discrete Hashing (LGDH), which simultaneously preserves the comprehensive manifold structure and discriminative balanced codes that are both constructed by label information into Hamming space. We develop a local category distribution of the nearest neighbors, to excavate the underlying manifold structure of heterogeneous data. To maximize the gaps of different categories, a balanced matrix is constructed by labels to generate hash codes with balanced bits. For multi-label data, we also design a novel multi-label manifold and balanced structure matrix to adapt the real-world scenarios. An effective discrete optimization method is used to optimize our proposed objective function instead of the relaxation one. Extensive experiments on three benchmark datasets verify the effectiveness of LGDH. The comparison results demonstrat that LGDH achieves about 2% and 3% improved to different cross-modal tasks on average.

The field of machine learning has recently undergone significant advancements, driven by the development of sophisticated algorithms and the availability of large datasets. Among the various paradigms of machine learning, supervised learning has traditionally dominated, relying heavily on labeled data to train models. However, acquiring and annotating large amounts of labeled data is both time-consuming and costly, leading to a growing interest in self-supervised learning (SSL). This promising approach leverages unlabeled data to learn useful representations, which can then be fine-tuned for specific tasks with minimal labeled data Self-supervised learning creates learning signals from the data itself, thus eliminating the need for manual labeling. By solving auxiliary tasks, such as predicting the rotation angle of an image or filling in missing parts of data, SSL methods can learn robust feature representations that are useful for a wide range of downstream tasks. This paradigm reduces dependency on labeled data and exploits the abundance of unlabeled data available in various domains. The significance of SSL lies in its potential to democratize machine learning, making it more accessible and applicable across different fields. From natural language processing to computer vision and beyond, self-supervised techniques have demonstrated remarkable success in improving model performance while reducing reliance on extensive labeled datasets. As such, SSL represents a critical step towards more efficient and scalable machine learning solutions. This paper explores the development and implementation of a novel self-supervised learning solution based on machine learning principles. The proposed method achieves an accuracy of 97.6%, a mean absolute error (MAE) of 0.403, and a root mean square error (RMSE) of 0.203. A comprehensive overview of the underlying methodologies, the challenges faced, and the potential applications of this approach is provided. By examining the theoretical foundations and practical implementations, this study aims to contribute to the growing body of research in SSL and highlight its transformative impact on the machine learning landscape.Self-Supervised Learning.

Hashing techniques for approximate nearest neighbor search (ANNS) encode data points into a set of short binary codes, while trying to preserve the neighborhood structure of the original data as much as possible. With the binary codes, the task of ANNS can be easily conducted over large-scale dataset, due to the high efficiency of pairwise comparison with the Hamming distance. Although binary codes have low computation and storage cost, the data are heavily compressed so that partial neighborhood structure information would be inevitably lost. To address this issue, we propose to introduce the k-nearest neighbors (k-NNs) in the original space into the Hamming space (i.e., associating a binary code with its original k-NNs) to enhance the effectiveness of existing hashing techniques with little overhead. Based on this idea, we develop a novel search scheme for hashing techniques namely neighborhood voting, i.e., each point retrieved by a query code will vote for its neighbors and itself, and the more voted, the better candidates. In this way, search in hashing is not simply the collision between codes (i.e., query code and candidate code), but also the collision between neighbors (i.e., neighbors of candidate points). The underlying assumption is that the true neighbors of a query point should be close to each other, while points with similar binary codes but seldom be the neighbors of other candidate points would be false positives. We introduce a novel data structure called aggregated hash table for implementing our idea and accelerating the online search process. Experimental results show that our search scheme can significantly improve the search effectiveness while having good efficiency over different existing hashing techniques.

Segmenting single cells is a necessary process for extracting quantitative data from biological microscopy imagery. The past decade has seen the advent of machine learning (ML) methods to aid in this process, the overwhelming majority of which fall under supervised learning (SL) which requires vast libraries of pre-processed, human-annotated labels to train the ML algorithms. Such SL pre-processing is labor intensive, can introduce bias, varies between end-users, and has yet to be shown capable of robust models to be effectively utilized throughout the greater cell biology community. Here, to address this pre-processing problem, we offer a self-supervised learning (SSL) approach that utilizes cellular motion between consecutive images to self-train a ML classifier, enabling cell and background segmentation without the need for adjustable parameters or curated imagery. By leveraging motion, we achieve accurate segmentation that trains itself directly on end-user data, is independent of optical modality, outperforms contemporary SL methods, and does so in a completely automated fashion—thus eliminating end-user variability and bias. To the best of our knowledge, this SSL algorithm represents a first of its kind effort and has appealing features that make it an ideal segmentation tool candidate for the broader cell biology research community.

Motion trajectory recognition is important for characterizing the moving property of an object. The speed and accuracy of trajectory recognition rely on a compact and discriminative feature representation, and the situations of varying rotation, scaling and translation has to be specially considered. In this paper we propose a novel feature extraction method for trajectories. Firstly a trajectory is represented by a proposed point context, which is a rotation-scale-translation (RST) invariant shape descriptor with a flexible tradeoff between computational complexity and discrimination, yet we prove that it is a complete shape descriptor. Secondly, the shape context is nonlinearly mapped to a subspace by kernel nonparametric discriminant analysis (KNDA) to get a compact feature representation, and thus a trajectory is projected to a single point in a low-dimensional feature space. Experimental results show that, the proposed trajectory feature shows encouraging improvement than state-of-art methods.

Contrastive learning is an important technique in the field of machine learning for learning data representations. In the field of self-supervised visual representation learning, the strategy of positive and negative sample selection is key to improving the efficiency and effectiveness of model learning. Traditional self-supervised learning methods often employ random sampling strategies to select positive and negative samples, but this approach may result in uneven sample quality when dealing with complex datasets, thereby affecting the learning outcomes. To alleviate this problem, this study is dedicated to exploring more effective strategies for positive and negative sample selection and processing to optimize the self-supervised learning process. To this end, we propose an improved method of self-supervised learning called contrastive learning with enhanced diversity. On the one hand, this method utilizes the weight parameters of the DINO pre-trained model to initialize the feature extraction network of SimCLR, providing more accurate calculations of feature similarity. On the other hand, by setting a threshold on the feature similarity matrix and penalizing (subtracting 0.5 from) similarity scores that do not exceed this threshold, we reduce the excessive impact of high similarity scores on model training, thereby helping the model better distinguish between positive and negative samples. In downstream image classification tasks, we conducted detailed evaluations of the improved model, specifically including fine-tuning and linear evaluation aspects. Experiments show that the proposed approach improves the performance of the loss functions and improves the accuracy of the proposed SimCLR model.

Deep learning has emerged as a powerful tool in various domains, revolutionising machine learning research. However, one persistent challenge is the scarcity of labelled training data, which hampers the performance and generalisation of deep learning models. To address this limitation, researchers have developed innovative methods to overcome data scarcity and enhance deep model learning capabilities. Two prevalent techniques that have gained significant attention are transfer learning and self-supervised learning. Transfer learning leverages knowledge learned from pre-training on a large-scale dataset, such as ImageNet, and applies it to a target task with limited labelled data. This approach allows models to benefit from the learned representations and effectively transfer knowledge to new tasks, resulting in improved learning performance and generalisation. On the other hand, self-supervised learning focuses on training models using pretext tasks that do not require manual annotation, allowing them to learn valuable representations from large amounts of unlabelled data. These learned representations can then be fine-tuned for downstream tasks, mitigating the need for extensive labelled data. In recent years, transfer and self-supervised learning have found applications in various fields, including medical image processing, video recognition, and natural language processing. These approaches have demonstrated remarkable achievements, enabling breakthroughs in areas such as disease diagnosis, object recognition, and language understanding. However, while these methods offer numerous advantages, they also have limitations. For example, transfer learning may face domain mismatch issues between the pre-training and target domains, while self-supervised learning requires careful design of pretext tasks to ensure meaningful representations. This review paper explores the recent applications of these pre-training methods in various fields within the past three years. It delves into the advantages and limitations of each approach, assesses the performance of models employing these techniques, and identifies potential directions for future research. By providing a comprehensive review of current pre-training methods, this article offers guidance for selecting the best technique for specific deep learning applications to address the data scarcity issue.

Deep hashing methods utilize an end-to-end framework to mutually learn feature representations and hash codes, thereby achieving a better retrieval performance. Traditional supervised hashing methods adopt handcrafted features for hashing function learning and then generate hash codes through classification and quantization. The lack of adaptability and independence of the quantization procedure leads to low retrieval accuracy of supervised hashing methods with handcrafted features in image retrieval. In this study, a non-relaxation deep hashing method for fast image retrieval is proposed. In this method, a differentiable host thresholding function is used to encourage hash-like codes to approach -1 or 1 non-linearly at the output of the convolutional neural, instead of the symbol function for quantization used in the traditional method. The output of the host thresholding function is directly used to compute the network training error, and a loss function is elaborately designed with the norm to constrain each bit of the hash-like code to be as binary as possible. Finally, a symbol function is added outside the trained network model to generate binary hash codes for image storage and retrieval in a low-dimensional binary space. Extensive experiments on two large-scale public datasets show that our method can effectively learn image features, generate accurate binary hash codes, and outperform state-of-the-art methods in terms of the mean average precision.

Hashing has been extensively utilized in cross-modal retrieval due to its high efficiency in handling large-scale, high-dimensional data. However, most existing cross-modal hashing methods operate as offline learning models, which learn hash codes in a batch-based manner and prove to be inefficient for streaming data. Recently, several online cross-modal hashing methods have been proposed to address the streaming data scenario. Nevertheless, these methods fail to fully leverage the semantic information and accurately optimize hashing in a discrete fashion. As a result, both the accuracy and efficiency of online cross-modal hashing methods are not ideal. To address these issues, this paper introduces the Semantic Embedding-based Online Cross-modal Hashing (SEOCH) method, which integrates semantic information exploitation and online learning into a unified framework. To exploit the semantic information, we map the semantic labels to a latent semantic space and construct a semantic similarity matrix to preserve the similarity between new data and existing data in the Hamming space. Moreover, we employ a discrete optimization strategy to enhance the efficiency of cross-modal retrieval for online hashing. Through extensive experiments on two publicly available multi-label datasets, we demonstrate the superiority of the SEOCH method.

In recent years, the information retrieval (IR) community has witnessed the first successful applications of deep neural network models to short-text matching and ad-hoc retrieval tasks. However, the two communities - focused on deep neural networks and on IR - have less in common when it comes to the choice of programming languages. Indri, an indexing framework popularly used by the IR community, is written in C++, while Torch, a popular machine learning library for deep learning, is written in the light-weight scripting language Lua. To bridge this gap, we introduce Luandri (pronounced laundry), a simple interface for exposing the search capabilities of Indri to Torch models implemented in Lua.

Deep learning methodologies have significantly advanced the fields of computer vision and machine learning, enhancing performance across various tasks like classification, regression, and detection. In remote sensing for Earth observation, deep neural networks have propelled state-of-the-art results. However, a major drawback is their dependence on large annotated datasets, necessitating extensive human effort, especially in specialized domains like medical imaging or remote sensing. To mitigate this reliance on annotations, several self-supervised representation learning techniques have emerged, aiming to learn unsupervised image representations applicable to downstream tasks such as image classification, object detection, or semantic segmentation. Consequently, self-supervised learning approaches have gained traction in remote sensing. This article surveys the foundational principles of various self-supervised methods, focusing on scene classification tasks. We elucidate key contributions, analyze experimental setups, and synthesize findings from each study. Furthermore, we conduct comprehensive experiments on two public scene classification datasets to evaluate and benchmark different self-supervised models

Precision agriculture requires accurate discrimination between winter crops and weeds, but there is a lack of annotated image data for winter cropping systems. This paper investigates a two-stage deep learning approach that integrates self-supervised feature learning with supervised fine-tuning for winter crop and weed image classification. A new winter crop and weed image dataset, WinterCropWeedDB, is proposed and used in this paper, which contains 1,136 high-resolution images of six winter crop species and four weed species collected from agricultural fields in central India. In the first stage of self-supervised learning, an EfficientNet-B3 model is pre-trained using a SimCLR-style self-supervised learning approach with an InfoNCE loss function (temperature τ = 0.5) on the images. The average contrastive loss value reduces from 2.0712 in the first iteration to 1.6835 at the end of pretraining. In the second stage of supervised fine-tuning, the pre-trained EfficientNet-B3 model is fine-tuned with a supervised classifier head on the images and tested on a single internal validation split (30%) of the dataset. The fine-tuned model reaches a maximum validation accuracy of 98.27%, with a macro-averaged F1-score of 0.98. Gradient-weighted class activation mapping (Grad-CAM) and Grad-CAM++ are used on the fine-tuned model to provide a qualitative visualization of the image regions that contribute to class predictions. The experiment outcomes demonstrate the viability of using self-supervised pretraining and supervised fine-tuning for the classification of winter crop and weed images on a region-specific dataset, while also emphasizing the importance of additional testing on independent test sets.