32584 research outputs found
Sort by
Identifying Red Sponges on ARMS Plates by Preprocessing Images using Histogram Equalization
No full textSponges play a vital role in marine ecosystems, being the only organisms capable of converting dissolved organic matter (DOM) into particulate organic matter (POM). They provide nutrients for coral reefs to thrive in oligotrophic waters. Autonomous reef monitoring structures (ARMS) are used to measure the biodiversity of coral reefs by simulating the complex cavities inside reef structures. Organisms settle on them and scientists can retrieve them after a period of time for analysis. Images are taken of ARMS plates after they are retrieved. Human analysis is unsuitable for the analysis of ARMS plates due to the huge number of images. Computer vision is leveraged to automate the process using machine learning models such as Mask R-CNN. A possible improvement is to preprocess the images using histogram equalization before training the model
AUTOMATED SPONGE SEGMENTATION WITH MASK R-CNN ON ARMS PLATE IMAGES FOR REEF MONITORING
Get PDFClimate change-driven ocean warming and acidification are disrupting the ecological balance of coral reefs. Notably, these oceanic conditions are undermining coral health and accelerating their decline which is favoring some sponge species in outcompeting them for dominance. Although functional, these altered reefs destabilize the reef architecture, hinder nutrient cycling, and support fewer marine species. Thus, monitoring the growth and abundance of various sponges and understanding their roles at different stages of ecological succession in coral reefs is vital. Autonomous reef monitoring structures (ARMS) are often used for this purpose, but manual taxonomic analysis using their images is time-consuming, inconsistent and not scalable. Here, we developed preliminary Mask R-CNN models using transfer learning approach for automating the process of identifying sponge Clathrinidae in ARMS plate images. The models trained were effective in identifying the target sponge across both seen and unseen images, showcasing its learning potential and generalizability even with a small training dataset. They also show how different training focus can lead to differences in their recall and precision performances. Overall, this work highlights the potential of Mask R-CNN based models to facilitate automated coral reef monitoring and guide their conservation and restoration strategies
Large Language Models for Bacterial Genomic Analysis
No full textIdentification of bacterial gene sequences with agricultural applications has the potential to transform agricultural biotechnology. These genes can be used in environmentally friendly pest control strategies. One such use case is identifying genes with potential insecticidal properties. With an increasing number of genomic information and decreasing numbers of available annotated sequences, finding new insecticidal genes has become more challenging.The traditional methods relying on sequence alignment and annotated databases are not effective in detecting functionally relevant genes lacking close homology to known cases. This project investigates the data-driven classification of genes by sequence modeling. This research is focused on learning DNA sequence motifs and transferring them to distinguish between insecticidal and non-insecticidal genes. The study exhibits that decision-making functional information may be obtained from DNA with state-of-the-art machine learning methodologies and that deep models are capable of generalization to low-resource environments
Clustering Organ Cell Types
No full textThe Human Cell Atlas (HCA) created a reference map of all human cells. My project uses the Tabula Sapiens dataset, developed under HCA and based on single-cell RNA sequencing data, to explore cell type and tissue diversity. I performed experiments using the Elbow method and a formula based on dataset observations to determine the number of clusters, then applied k-means clustering on two representative subsets of the All Cells dataset. Clusters were selected for analysis using Shannon’s Diversity Index and Pielou’s Evenness. A novel algorithm based on the cell differentiation tree was used to validate the biological coherence of the clusters. Cell type emerged as the most informative feature for interpreting cluster structure. The report concludes with a summary of results and recommendations for future research
Crafting Research That Resonates
Get PDFA short presentation made to the students of SJSU’s Gateway PhD Program with Manchester Metropolitan University on August 5, 2025, at their annual research workshop. It serves as some general advice and inspiration to aspiring doctoral students.
More about the program here: https://ischool.sjsu.edu/gateway-phd-library-and-information-managementhttps://scholarworks.sjsu.edu/oer/1015/thumbnail.jp
Teacher Workplace Bullying: Victim and Bystander Perspectives
Get PDFWorkplace bullying is a phenomenon that is addressed within a broad context. This study focuses on teacher experiences of bullying. The focus will examine the impacts horizontal and vertical bullying can have on teacher victims and bystanders within K-12th within California. The purpose of this study is to give administrators awareness of these issues from the victim and bystanders’ lens. The study implemented a mixed methods design to give insight to teacher perceptions on the impacts it can have on their teacher efficacy, teacher retention, and their emotional well-being. By looking at how often PK-12th grade teachers experience workplace bullying, this dissertation will inform future school policies on bullying
Spartan Daily, August 20, 2025
Get PDFVolume 165, Issue 1https://scholarworks.sjsu.edu/spartan_daily_2025/1043/thumbnail.jp
The Sketches of Infinite Data and Algorithms for Real-Time Data Insights
Get PDFHow are machine learning algorithms able to answer questions from any nook and corner of the World Wide Web? How are trending hashtags from the near infinite microblog posts, unique visitors and other distinct counts in the near infinite website traffic determined? How do blogging websites avoid recommending articles a user has previously read? In general, how can we answer complex queries about enormous data streams without storing them entirely, in real-time? The answer often lies in clever approximation algorithms and data sketches that capture essential properties using vastly reduced space. The relentless flow of data in modern systems indeed presents significant challenges. These data streams are often too large to store and too fast to process exhaustively with traditional methods. This talk introduces key sketching and approximation techniques that help generate real-time data insights by processing data streams.https://scholarworks.sjsu.edu/oer/1016/thumbnail.jp
Comparative Analysis of Forensic Science Systems: Scotland vs. America
Get PDFForensic science plays an important role in the justice system of both the United States and Scotland, but the structure, application, and effectiveness differ significantly. This research paper presents a comparative analysis of forensic science in two different legal and institutional contexts, examining differences between the two forensic science systems. This includes forensic laboratories and services, education and training, policing, court procedures, legal framework, and crime scene investigation. The analysis highlights significant differences such as Scotland’s centralized system versus America’s decentralized system, which impacts resource allocation and standardization for forensic science. Despite major differences, both countries emphasize scientific reliability, standardized forensic methodologies, and professional development