Retrieval Domain Research Articles

Plant Leaf Disease Detection Using Integrated Color and Texture Features

In the realm of precision agriculture, a pivotal challenge lies in the detection, identification, and grading of crop diseases. This multifaceted task necessitates the involvement of expert human resources and time-sensitive actions aimed at mitigating the risks of production losses and the rapid spread of diseases. The effectiveness of the majority of developed systems in this domain hinges on the quality of image features and disease segmentation accuracy. This paper presents a comprehensive research endeavor in the domain of Content-Based Image Retrieval (CBIR), specifically tailored to detect and classify leaf diseases. The proposed system integrates both color and texture features to underpin its functionality, providing a robust framework for accurate disease detection. By leveraging advanced image processing techniques, the system enhances the precision of disease identification, which is crucial for timely and effective intervention in agricultural practices. To evaluate the system’s performance, maize leaves afflicted by rust and blight serve as prime candidates for testing. These diseases were chosen due to their prevalence and significant impact on crop yield. The experimental results demonstrate that the developed system consistently excels in its disease detection and identification tasks, boasting an impressive efficiency rate of 98.33%. This high level of accuracy underscores the potential of the system to be a valuable tool in precision agriculture, aiding farmers and agricultural experts in maintaining healthy crops and optimizing production. The integration of color and texture features not only improves the detection accuracy but also provides a comprehensive understanding of the disease characteristics. This dual-feature approach ensures that the system can distinguish between different types of diseases with high precision, making it a versatile solution for various agricultural applications. The findings of this research highlight the importance of advanced image analysis techniques in enhancing the capabilities of disease detection systems, paving the way for more efficient and effective agricultural practices.

Semantic decomposition and enhancement hashing for deep cross-modal retrieval

Deep hashing has garnered considerable interest and has shown impressive performance in the domain of retrieval. However, the majority of the current hashing techniques rely solely on binary similarity evaluation criteria to assess the semantic relationships between multi-label instances, which presents a challenge in overcoming the feature gap across various modalities. In this paper, we propose semantic decomposition and enhancement hashing (SDEH) by extensively exploring the multi-label semantic information shared by different modalities for cross-modal retrieval. Specifically, we first introduce two independent attention-based feature learning subnetworks to capture the modality-specific features with both global and local details. Subsequently, we exploit the semantic features from multi-label vectors by decomposing the shared semantic information among multi-modal features such that the associations of different modalities can be established. Finally, we jointly learn the common hash code representations of multimodal information under the guidelines of quadruple losses, making the hash codes informative while simultaneously preserving multilevel semantic relationships and feature distribution consistency. Comprehensive experiments on four commonly used multimodal datasets offer strong support for the exceptional effectiveness of our proposed SDEH.

Performance Evaluation in Multimedia Retrieval

Performance evaluation in multimedia retrieval, as in the information retrieval domain at large, relies heavily on retrieval experiments, employing a broad range of techniques and metrics. These can involve human-in-the-loop and machine-only settings for the retrieval process itself and the subsequent verification of results. Such experiments can be elaborate and use-case-specific, which can make them difficult to compare or replicate. In this paper, we present a formal model to express all relevant aspects of such retrieval experiments, as well as a flexible open-source evaluation infrastructure that implements the model. These contributions intend to make a step towards lowering the hurdles for conducting retrieval experiments and improving their reproducibility.

GPT vs. Google: A Comparative Study of Self-Code Learning in ODL Students

This research investigates the comparative effectiveness of the ChatGPT and the Google search engine in facilitating the self-learning of JavaScript functions among undergraduate open and distance learning students. Employing a quasi-experimental post-test control group design, the study analyzed variables of disorientation, satisfaction, cognitive load, navigation performance, and test scores among 471 learners. Results indicated a marked advantage for ChatGPT users in reducing cognitive load and disorientation, improving navigation performance and increasing test scores. The study also explores the evolving dynamics between conventional keyword-based search engines and AI-driven conversational models, reflecting on their mutual influence in the domain of information retrieval. These findings suggest that AI chatbots like ChatGPT not only enhance the learning experience but also signify a paradigm shift in how information is accessed and processed in educational settings. The research contributes to the understanding of AI applications in education, particularly in self-directed learning scenarios in open and distance learning.

Impact of corner radius of the aperture on the accuracy of phase retrieval analysis in single-frame CDI for nanomaterials

Numerous efforts have been developed to advance the process of Ptychography, leading to the emergence of a phenomenon known as single-frame Coherent X-ray Diffraction Imaging (CDI). It represents an advanced form of Ptychography in which a well-defined coherent X-ray beam is directed onto the sample to generate a diffraction image that includes intensity information but lacks phase details. However, phase retrieval in single-frame CDI faces challenges like overlapping loss, inadequate samples, or phase ambiguity due to the characteristics of this method. The optimising effort has been conducted in the form of aperture implementation to address these problems, particularly in the iterative phase retrieval domain. This study explores non-point symmetrical apertures, specifically triangles, to enhance image reconstruction quality in iterative phase retrieval. We developed a simulation methodology to explore the impact of corner radius on the apertures in performing single-frame CDI. Our simulator generates diffraction images for each shape, which are processed through iterative algorithms and deep learning to reconstruct images. The effectiveness of each shape is assessed by visualising the loss function results, providing insights into optimal aperture designs for diffraction imaging, and contributing to the resolution of phase retrieval challenges in single-frame CDI.

Music Emotion Classification: A Literature Review

Music, as an art form, combines rhythm and sound to form a functional melodic structure, uniquely capable of conveying emotions non-verbally. Within the domain of Music Information Retrieval (MIR), Music Emotion Classification (MEC) represents a specialized subset dedicated to the identification and labeling of emotional attributes in songs. This is achieved by extracting and comparing features from musical compositions. This research aims to discern the contemporary landscape of research and its associated research gaps in this domain. The study comprised a collection of publications found through searches conducted on Google Scholar between the years 2006 and 2023, with the search terms: Music Emotion Classification, Music Emotion Classification in Sri Lanka, Music Emotion Recognition, and Emotion Classification in Music. This study was narrowed down to the research that utilized audio files for classification. Among the initial set of 42 studies, 20 were selected for detailed analysis using the purposive sampling method. The review encompassed essential aspects, including acoustic feature analysis, emotional modeling, classification methodologies, and performance evaluation. The findings highlighted a paucity of research considering cultural, regional, and linguistic variations. The most often used acoustic features encompassed rhythm, pitch, timbre, spectral, and harmony whereas the most frequently used emotion categories for the classification were happiness, anger, sadness, and relaxation. Support Vector Machine (SVM) was the most used machine learning algorithm for classification, although regression methods, neural network-based approaches, and fuzzy classifications were also explored. Notably, the adoption of multi-modal approaches for emotion classification, as well as multi-labeled emotion classification, remained limited. These insights underscore the need for future research to address the cultural and language diversity of datasets, explore innovative classification techniques, and embrace multi-modal and multi-labeled emotional classification methodologies in the context of music emotion classification within MIR.

Assembly and fission of tubular carriers mediating protein sorting in endosomes.

Endosomes are central protein-sorting stations at the crossroads of numerous membrane trafficking pathways in all eukaryotes. They have a key role in protein homeostasis and cellular signalling and are involved in the pathogenesis of numerous diseases. Endosome-associated protein assemblies or coats collect transmembrane cargo proteins and concentrate them into retrieval domains. These domains can extend into tubular carriers, which then pinch off from the endosomal membrane and deliver the cargoes to appropriate subcellular compartments. Here we discuss novel insights into the structure of a number of tubular membrane coats that mediate the recruitment of cargoes into these carriers, focusing on sorting nexin-based coats such as Retromer, Commander and ESCPE-1. We summarize current and emerging views of how selective tubular endosomal carriers form and detach from endosomes by fission, highlighting structural aspects, conceptual challenges and open questions.

An empirical study of excitation and aggregation design adaptions in CLIP4Clip for video–text retrieval

CLIP4Clip model transferred from the CLIP has been the de-factor standard to solve the video clip retrieval task from frame-level input, triggering the surge of CLIP4Clip-based models in the video–text retrieval domain. In this work, we rethink the inherent limitation of widely-used mean pooling operation in the frame features aggregation and investigate the adaptions of excitation and aggregation design for discriminative video representation generation. We present a novel excitation-and-aggregation design, including (1) The excitation module is available for capturing non-mutually-exclusive relationships among frame features and achieving frame-wise features recalibration, and (2) The aggregation module is applied to learn exclusiveness used for frame representations aggregation. Similarly, we employ the cascade of sequential module and aggregation design to generate discriminative video representation in the sequential type. Besides, we adopt the excitation design in the tight type to obtain representative frame features for multi-modal interaction. The proposed modules are evaluated on three benchmark datasets of MSR-VTT, ActivityNet and DiDeMo, achieving MSR-VTT (43.9 R@1), ActivityNet (44.1 R@1) and DiDeMo (31.0 R@1). They outperform the CLIP4Clip results by +1.2% (+0.5%), +4.5% (+1.9%) and +9.5% (+2.7%) relative (absolute) improvements, demonstrating the superiority of our proposed excitation and aggregation designs. We hope our work will serve as an alternative for frame representations aggregation and facilitate future research.

Accurate Segmentation Method for Roadside Lampposts Based on Vehicle-Mounted Lidar Point Clouds

Abstract. The extraction of roadside lampposts constitutes a significant research focus within the domain of vehicular LiDAR point cloud object retrieval. Addressing the complexities inherent in discerning lampposts amidst convoluted vegetation in diverse roadway settings, this study introduces an innovative, progressive technique for the individualized extraction of lampposts utilizing vehicular LiDAR point clouds. The proposed method initiates with a bifurcation of the primary point cloud into terrestrial and aerial subsets via the Cloth Simulation Filter (CSF) algorithm. Subsequent processes involve the extraction of distinct lamppost structures from aerial point clouds through a methodology integrating elevation-normalized spatial partitioning and directional coverage analysis, thereby facilitating precise lamppost localization. The culmination of this process involves the refinement of lamppost point clouds through supervoxel clustering complemented by a non-discretization filter grounded in Principal Component Analysis (PCA). The efficacy of this novel approach is substantiated through empirical studies employing a LiDAR dataset encompassing extensive adhesive scenarios, whereupon comparative analysis with extant methodologies reveals its enhanced proficiency in isolating individual lampposts in complex environments.

A Precise Framework for Rice Leaf Disease Image-Text Retrieval Using FHTW-Net.

Cross-modal retrieval for rice leaf diseases is crucial for prevention, providing agricultural experts with data-driven decision support to address disease threats and safeguard rice production. To overcome the limitations of current crop leaf disease retrieval frameworks, we focused on four common rice leaf diseases and established the first cross-modal rice leaf disease retrieval dataset (CRLDRD). We introduced cross-modal retrieval to the domain of rice leaf disease retrieval and introduced FHTW-Net, a framework for rice leaf disease image-text retrieval. To address the challenge of matching diverse image categories with complex text descriptions during the retrieval process, we initially employed ViT and BERT to extract fine-grained image and text feature sequences enriched with contextual information. Subsequently, two-way mixed self-attention (TMS) was introduced to enhance both image and text feature sequences, with the aim of uncovering important semantic information in both modalities. Then, we developed false-negative elimination-hard negative mining (FNE-HNM) strategy to facilitate in-depth exploration of semantic connections between different modalities. This strategy aids in selecting challenging negative samples for elimination to constrain the model within the triplet loss function. Finally, we introduced warm-up bat algorithm (WBA) for learning rate optimization, which improves the model's convergence speed and accuracy. Experimental results demonstrated that FHTW-Net outperforms state-of-the-art models. In image-to-text retrieval, it achieved R@1, R@5, and R@10 accuracies of 83.5%, 92%, and 94%, respectively, while in text-to-image retrieval, it achieved accuracies of 82.5%, 98%, and 98.5%, respectively. FHTW-Net offers advanced technical support and algorithmic guidance for cross-modal retrieval of rice leaf diseases.

A Learn-to-Rank Approach to Medicine Selection for Patient Treatments

Aim/Purpose: This research utilized a learn-to-rank algorithm to provide medical recommendations to prescribers. The algorithm has been utilized in other domains, such as information retrieval and recommender systems. Background: Ranking the possible medical treatments according to diagnoses of the medical cases is very beneficial for doctors, especially during the coding process. Methodology: We developed two deep learning pointwise learn-to-rank models within one prediction pipeline: one for predicting the top possible active ingredients from disease features, the other for ranking actual medicines codes from diseases and the ingredients features. Contribution: A new learn-to-rank deep learning model has been developed to rank medical procedures based on datasets collected from insurance companies. Findings: We ran 18 cross-validation trials on a confidential dataset from an insurance company. We obtained an average normalized discounted cumulative gain (NDCG@8) of 74% with a 5% standard deviation as a result of all 18 experiments. Our approach outperformed a known approach used in the information retrieval domain in which data is represented in LibSVM format. Then, we ran the same trials using three learn-to-rank models – pointwise, pairwise, and listwise – which yielded average NDCG@8 of 71%, 72%, and 72%, respectively. Recommendations for Practitioners: The proposed model provides an insightful approach to helping to manage the patient’s treatment process. Recommendation for Researchers: This research lays the groundwork for exploring various applications of data science techniques and machine learning algorithms in the medical field. Future studies should focus on the significant potential of learn-to-rank algorithms across different medical domains, including their use in cost-effectiveness models. Emphasizing these algorithms could enhance decision-making processes and optimize resource allocation in healthcare settings. Impact on Society: This will help insurance companies and end users reduce the cost associated with patient treatment. It also helps doctors to choose the best procedure and medicines for their patients. Future Research: Future research is required to investigate the impact of medicine data at a granular level.

Fast metric multi-view hashing for multimedia retrieval

The acquisition of multi-view hash representation for heterogeneous data holds paramount importance in the domain of multimedia retrieval. The limited retrieval precision observed in current approaches stems from their inadequate integration of multi-view features and their failure to effectively leverage the metric information available from diverse samples. Commonly employed fusion methods, such as concatenation or weighted sum, are insufficient in capturing the complementarity among multiple view features. Furthermore, these methods neglect the valuable information contributed by dissimilar samples. To address these challenges, we propose an innovative method termed Fast Metric Multi-View Hashing (FMMVH). Our approach showcases the superiority of gate-based fusion over traditional methods, as substantiated by extensive empirical evidence. Additionally, this paper proposes a novel deep metric loss function to enable the utilization of metric information from dissimilar samples. We exclusively train our method using this single loss function. To enhance practical applicability in industrial production environments, we employ model compression techniques to optimize the proposed method. On benchmark datasets such as MIR-Flickr25K, NUS-WIDE, and MS COCO, the performance of our FMMVH method significantly surpasses that of existing state-of-the-art methods, demonstrating improvements of up to 7.47% in mean Average Precision (mAP).

Tailoring ontology retrieval for supporting requirements analysis

It is well accepted that domain ontologies can support requirement analysis activities, particularly in detecting inconsistencies and incompleteness of requirement models. These benefits critically depend on the provision of a suitable ontology. We observe the context of supporting requirement analysis provides both opportunities and restrictions when choosing the most appropriate ontology retrieval mechanisms. Requirement models are the basis for retrieving the most influential ontologies and are not the typical retrieval domain ontologies. For instance, a retrieval ontology derived from the requirement is not expected only to be a hierarchical taxonomy, nor is it limited to the boundaries of a single domain, nor does it cover any particular domain completely. Hence, retrieval methods cannot be based on classes only and computational constraints do not necessarily apply as the retrieval process is expected to run only once at the outset of the analysis phase. It is also important to assume that the retrieval in this context is targeting multiple ontologies describing multiple but related domains. In this paper, we deduce that avoiding structural based retrieval mechanisms in fact benefits to the requirement models. Instead, we formulate a new retrieval method based on the PageRank algorithm that takes into account the indirect influences of various concepts within plausible supporting ontologies. This paper provides an empirical analysis that evidences the strength of our retrieval algorithm in supporting the identification of ontologies to support requirement analysis.

Multi-Modal Medical Image Matching Based on Multi-Task Learning and Semantic-Enhanced Cross-Modal Retrieval

With the continuous advancement of medical imaging technology, a vast amount of multi-modal medical image data has been extensively utilized for disease diagnosis, treatment, and research. Effective management and utilization of these data becomes a pivotal challenge, particularly when undertaking image matching and retrieval. Although numerous methods for medical image matching and retrieval exist, they primarily rely on traditional image processing techniques, often limited to manual feature extraction and singular modality handling. To address these limitations, this study introduces an algorithm for medical image matching grounded in multi-task learning, further investigating a semantic-enhanced technique for cross-modal medical image retrieval. By deeply exploring complementary semantic information between different modality medical images, these methods offer novel perspectives and tools for the domain of medical image matching and retrieval.

A rule-based obfuscating focused crawler in the audio retrieval domain

The detection of violations of intellectual properties on multimedia files is a critical problem for the current infrastructure of the Internet, especially within very large document collections. To contrast such a problem, either proactive or reactive methods are used. The first category prevents the upload of infringing files themselves by comparing illegal files with a reference collection, while the second one responds to reports made by third parties or artificial intelligence systems in order to delete files deemed illegal. In this article we propose an approach that is both reactive and proactive at the same time, with the aim of preventing the deletion of legal uploads of files (or modifications of such files, such as remixes, parodies and other edits) due to the presence of illegal uploads on a platform. We developed a rule-based obfuscating focused crawler able to work with audio files in the Audio Information Retrieval (AIR) domain, but its use can be easily extended to other multimedia file types, such as videos or textual documents. Our proposed model automatically scans multimedia files uploaded to the public collection only when a user query is submitted to it. We will also show experimental results obtained during tests on a known musical collection. Several combinations of specific Neural Network-Similarity Scorer solutions are shown, and we will discuss the strength and efficiency of each combination.

An evaluation metric for image retrieval systems, using entropy for grouped precision of relevant retrievals

Evaluating the performance of Content-Based Image Retrieval (CBIR) systems is a challenging and intricate task, even for experts in the field. The literature presents a vast array of CBIR systems, each applied to various image databases. Traditionally, automatic metrics employed for CBIR evaluation have been borrowed from the Text Retrieval (TR) domain, primarily precision and recall metrics. However, this paper introduces a novel quantitative metric specifically designed to address the unique characteristics of CBIR. The proposed metric revolves around the concept of grouping relevant images and utilizes the entropy of the retrieved relevant images. Grouping together relevant images holds great value from a user perspective, as it enables more coherent and meaningful results. Consequently, the metric effectively captures and incorporates the grouping of the most relevant outcomes, making it highly advantageous for CBIR evaluation. Additionally, the proposed CBIR metric excels in differentiating between results that might appear similar when assessed using other metrics. It exhibits a superior ability to discern subtle distinctions among retrieval outcomes. This enhanced discriminatory power is a significant advantage of the proposed metric. Furthermore, the proposed performance metric is designed to be straightforward to comprehend and implement. Its simplicity and ease of use contribute to its practicality for researchers and practitioners in the field of CBIR. To validate the effectiveness of our metric, we conducted a comprehensive comparative study involving prominent and well-established CBIR evaluation metrics. The results of this study demonstrate that our proposed metric exhibits robust discrimination power, outperforming existing metrics in accurately evaluating CBIR system performance.

SketchANIMAR: Sketch-based 3D animal fine-grained retrieval

The retrieval of 3D objects has gained significant importance in recent years due to its broad range of applications in computer vision, computer graphics, virtual reality, and augmented reality. However, the retrieval of 3D objects presents significant challenges due to the intricate nature of 3D models, which can vary in shape, size, and texture, and have numerous polygons and vertices. To this end, we introduce a novel SHREC challenge track that focuses on retrieving relevant 3D animal models from a dataset using sketch queries and expedites accessing 3D models through available sketches. Furthermore, a new dataset named ANIMAR was constructed in this study, comprising a collection of 711 unique 3D animal models and 140 corresponding sketch queries. Our contest requires participants to retrieve 3D models based on complex and detailed sketches. We receive satisfactory results from eight teams and 204 runs. Although further improvement is necessary, the proposed task has the potential to incentivize additional research in the domain of 3D object retrieval, potentially yielding benefits for a wide range of applications. We also provide insights into potential areas of future research, such as improving techniques for feature extraction and matching and creating more diverse datasets to evaluate retrieval performance.

Building a search tool for compositely annotated entities using Transformer-based approach: Case study in Biosimulation Model Search Engine (BMSE).

The Transformer-based approaches to solving natural language processing (NLP) tasks such as BERT and GPT are gaining popularity due to their ability to achieve high performance. These approaches benefit from using enormous data sizes to create pre-trained models and the ability to understand the context of words in a sentence. Their use in the information retrieval domain is thought to increase effectiveness and efficiency. This paper demonstrates a BERT-based method (CASBERT) implementation to build a search tool over data annotated compositely using ontologies. The data was a collection of biosimulation models written using the CellML standard in the Physiome Model Repository (PMR). A biosimulation model structurally consists of basic entities of constants and variables that construct higher-level entities such as components, reactions, and the model. Finding these entities specific to their level is beneficial for various purposes regarding variable reuse, experiment setup, and model audit. Initially, we created embeddings representing compositely-annotated entities for constant and variable search (lowest level entity). Then, these low-level entity embeddings were vertically and efficiently combined to create higher-level entity embeddings to search components, models, images, and simulation setups. Our approach was general, so it can be used to create search tools with other data semantically annotated with ontologies - biosimulation models encoded in the SBML format, for example. Our tool is named Biosimulation Model Search Engine (BMSE).

Recent Query Reformulation Approaches for Information Retrieval System - A Survey

Abstract:Around trillions of data are uploaded to the internet every year. Extracting useful information using only a few keywords has become a major challenge. The field of Query Reformulation (QR) is focused on the efficient retrieval of information to overcome this. It is widely used in the domain of information retrieval (IR) and related fields such as search engines, multimedia IR, cross-language IR, recommender systems, and so on. Query reformulation techniques incur extra computational costs. Due to this reason, the use of query reformulation techniques is sometimes prohibited in internet searches as searching over the internet requires a fast response time. But due to the success of NLP (Natural Language Processing) using machine learning/deep learning in recent years, there has been a boom of study in this area. In this literature, a variety of term selection, term extraction, and query reformulation strategies based on recent technologies used by researchers have been presented, necessitating a wide survey to focus research in this promising area. Recent QR approaches and the datasets, techniques, and evaluation metrics used in this paper will help researchers understand and focus more on research in this promising area so that a better solution will be proposed. From the survey, it may be observed that one of the hottest subjects in the field of IR right now is applying deep learning to IR systems for query reformulation.

Report on the 1st Training School on Domain Specific Systems for Information Extraction and Retrieval (DoSSIER 2022)

The DoSSIER project, an European Training Network, was kicked-off in late 2019. The first PhD candidate recruiting activities took place at the beginning of the Covid-19 pandemic and the first years of the project lived in the on-line universe. Naturally, this posed certain challenges to the cohesion of our network. With some delay due to the lock-downs and travel restrictions, our first Training School was successfully organized in September 2022 and it constituted one of the few occasions we had where the members of DoSSIER network met in person. This report gives an account on the first Summer School organized by DoSSIER, for the DoSSIER Early Stage Researchers, a school set in the picturesque village of Olympiada, which gave amply opportunity of interaction between students and senior researchers in the Information Retrieval domain. Date: 25--30 September, 2022. Website: https://dossier-project.eu/1st-dossier-training-school.