Centrum Wiskunde & Informatica’s Post

🗞️ 𝐀 𝐰𝐢𝐥𝐝 𝐩𝐚𝐩𝐞𝐫 𝐡𝐚𝐬 𝐚𝐩𝐩𝐞𝐚𝐫𝐞𝐝 ✨! Last year I interned at Naturalis Biodiversity Center Nederland and Centrum Wiskunde & Informatica which has resulted in the paper ForametCeTera, published yesterday (open access). With Tristan van Leeuwen and Willem Renema, I researched how we can speed up foraminifera analysis. These 𝑖𝑛𝑐𝑟𝑒𝑑𝑖𝑏𝑙𝑦 𝑡𝑖𝑛𝑦 organisms tell us a lot about the destructive effects of climate change. But! You need to analyse them one-by-one… by hand… 🙃🥱 So, research objective: automate that using computer vision and deep learning! Of course, this requires data. We created a high-throughput CT scanning workflow (powered by coffee creamer™) and scanned some 500 specimens to create ForametCeTera, a first of its kind dataset. This data can be used... 🤖 To directly train neural models 🧱 As digital Lego to create synthetic datasets to create neural models with 📈 To expand an ever growing digital foraminifera database Many thanks to Willem Renema, Tristan van Leeuwen and Bertie Joan van Heuven for your guidance. Also thanks to Alexander Skorikov, Richard Schoonhoven, PhD and Dirk Schut at CWI for their assistance!   The paper’s full title is “ForametCeTera, a novel CT scan dataset to expedite classification research of (non-)foraminifera”. All data and code is open access and available through the article website. https://lnkd.in/eECZBmtn   #climatechange #scientificpublication #machinelearning #computervision #deeplearning #biodiversity #biodiversiteit #foraminifera

I'm thrilled to announce our paper, titled 'Early Detection of Bark Beetle Attack Using Remote Sensing and Machine Learning: A Review,' published in the prestigious ACM Computing Surveys (IF: 16.6, ranked 3/111 in Computer Science Theory & Methods). I extend my sincere gratitude to my collaborators: Prof. Nadir Erbilgin and Prof. Nilanjan Ray from the University of Alberta, and Dr. Devin Goodsman from Natural Resources Canada | Ressources naturelles Canada for their valuable contributions to our research. https://lnkd.in/gz4nDq76 This review goes beyond previous efforts, encompassing all remote sensing systems and focusing on ML/DL methods, highlighting their strengths and weaknesses. In this review, we meticulously analyze the existing literature, focusing on multi- and hyper-spectral analyses. We distill knowledge on various aspects, including bark beetle species and attack phases, with a primary emphasis on early-stage detection, host trees, study regions, remote sensing platforms and sensors, spectral/spatial/temporal resolutions, spectral signatures, spectral vegetation indices, ML approaches, learning schemes, task categories, models, algorithms, classes/clusters, features, and DL networks and architectures. While our research highlights the promise of DL methods and the RF algorithm in detecting subtle changes across visible, thermal, and short-wave infrared spectral regions, it's important to note that challenges persist. Our paper not only sheds light on these challenges but also explores the myriad of opportunities, offering a comprehensive understanding of the current research landscape and paving the way for future directions in this fascinating field. ACM, Association for Computing Machinery #Research #MachineLearning #RemoteSensing #DeepLearning #uofa

Hello everyone, I'm pleased to announce that our research paper titled "Multiple Seed Segregation Using Image Processing" is now live on IEEE Xplore! We first presented it at the "2023 International Conference on Network, Multimedia and Information Technology (NMITCON)" in September"23. The paper discusses about how the #deeplearning algorithm of "YOLOv5" can be used to #segregate good and bad #seeds from #multiple variety of mix seeds along with it's seed type. I would like to extend my sincerest gratitude and appreciation to my co-authors; Saquib Akhtar and Mohammed Mehtab Alam for all their co-ordination and support as well as my mentor Deepali Koppad for her invaluable advice and guidance. At last, I also want to appreciate the efforts of Suma K V and Nethra Nagarajappa Without all this paper would never have been possible. It was a great experience writing the paper and hope to publish more such papers in future. You can find the paper on IEEE Xplore- https://lnkd.in/gCqng4ft #research #machinelearning #deeplearning #ai #aicommunity #seeds #agriculture #ieee #conference #engineering #connections

For another hypothetical concept, let's combine ideas from "The Quest for Simple Rules to Build a Microbial Community," "Perfecting the Art of Sensible Nonsense" in cryptography, and "In Mathematics, It Often Takes a Good Map to Find Answers." This blend leads to a concept I'll call **Cryptographic Ecosystem Mapping (CEM)**. ### Concept: Cryptographic Ecosystem Mapping (CEM) **Overview:** CEM aims to model complex ecosystems using cryptographic principles, providing a novel way to understand and simulate the interactions within biological communities. It also incorporates mathematical mapping techniques to navigate these complex systems. **Hypothetical Formula/Algorithm:** 1. **Ecosystem Modeling Using Cryptographic Functions:** - Each species in the ecosystem is represented by a unique cryptographic function or hash. - Interactions between species (like predation, symbiosis) are modeled as cryptographic operations that combine different species' hashes. 2. **Dynamic System Mapping:** - A mathematical mapping technique is employed to chart these interactions, akin to creating a 'map' of the ecosystem based on cryptographic interactions. - The map evolves as species interact, reflecting the dynamic nature of ecosystems. 3. **Ecosystem Interaction Rules:** - \( \text{Interaction}_{species1, species2} = \text{CryptoFunction}(\text{Hash}_{species1}, \text{Hash}_{species2}) \) - Where \( \text{CryptoFunction} \) denotes the cryptographic operation that models the interaction. 4. **Navigating the Ecosystem Map:** - Utilize advanced mathematical techniques, like those in topology or graph theory, to navigate and analyze the cryptographic ecosystem map. - This can help identify key species, interaction hubs, and vulnerabilities in the ecosystem. **Applications:** - CEM could be a powerful tool in conservation biology, helping to understand and predict the impacts of changes (like species extinction) in an ecosystem. - It could also find use in ecological informatics, aiding in the design of sustainable ecosystems, whether natural or in controlled environments like agriculture. **Challenges and Further Development:** - Defining appropriate cryptographic functions that accurately represent biological interactions. - Ensuring the ecosystem map remains analyzable and meaningful as the complexity of interactions grows. - Balancing the model's complexity with real-world applicability. Again, this is a speculative idea meant to demonstrate how interdisciplinary approaches can lead to novel concepts. Actual implementation would require extensive research and collaboration across biology, cryptography, and mathematics.

Breaking Barriers in Source-Free Domain Adaptation: NOTELA’s Impact on Bioacoustics and Vision Domains Quick Read: https://lnkd.in/g_ARR4we Paper: https://lnkd.in/gf9Qt_MW #artificialintelligence #machinelearning

🤔 Wondering what the link is between #reproducible #bioimaging and #deep #space #communication 🚀 ? Check out the EMBO commentary by our VIB Technologies colleagues at Tech Training (Alexander Botzki), Bioimaging Core Leuven (Sebastian Munck, Tatiana Woller) and Data Core (Tatiana Woller) 👇

📢 Exciting Publication Alert! 📢 I am thrilled to announce that my latest research paper, titled ‘’A Novel Transfer Learning Approach for Detection of Pomegranates Growth Stages", has been published in the IEEE Access journal. This work explores innovative methods to leverage transfer learning for accurately detecting the various growth stages of pomegranates, showcasing the potential of AI in enhancing agricultural practices. You can read the full paper here: [DOI: 10.1109/ACCESS.2024.3365356](https://lnkd.in/dKsXFYkH). Thank you for your support and interest in my research journey. Stay tuned for more exciting updates! #Research #Publication #IEEE #ArtificialIntelligence #MachineLearning #Agriculture #TransferLearning #AI #Pomegranates

Publication Alert!! I am happy to share the first Publication of 2024 in Q1 Journal "Bulletin of Engineering Geology and the Environment" Title - Hybrid soft computing models for predicting unconfined compressive strength of lime stabilized soil using strength property of virgin cohesive soil Abstract - This work introduces an optimal performance model for predicting the unconfined compressive strength (UCS) of lime-stabilized soil using the machine (ensemble tree (ET), Gaussian process regression (GPR), and decision tree (DT), support vector machine (SVM)), and hybrid (relevance vector machine (RVM)) learning computational techniques. The conventional (non-optimized) and hybrid (genetic (GA) and particle swarm algorithm optimized (PSO)) RVM models have been developed and compared with machine learning models. For the first time, UCS of virgin cohesive soil has been used as input variable to predict the UCS of lime-stabilized soil. A database of 371 results of lime-stabilized soil has been compiled from the literature and used to create training, testing, and validation databases. Furthermore, the multicollinearity levels for each input variable, i.e., lime content, UCS of cohesive soil, and curing period, have been determined as weak for the overall database. The performance of built-in models has been measured by three new index performance metrics: the a20-index, the index of scatter (IOS), and the index of agreement (IOA). This research concludes that the weak multicollinearity of input variables affects the performance of the non-optimized RVM models. Also, the ensemble tree has performed better than SVM, DT, and GPR because it consists of the number of trees. The overall performance comparison concludes that the PSO-optimized Laplacian kernel–based RVM model UCS16 outperformed all models with higher a20-index (testing = 67.30, validation = 55.95), IOA (testing = 0.8634, validation = 0.7795), and IOS (testing = 0.2799, validation = 0.3506) and has been recognized as the optimal performance model. ANOVA, Z, and Anderson-darling tests reject the null hypothesis for the present research. The lime content influences the prediction of UCS of lime-stabilized soil. The computational cost and external validation results show the robustness of model UCS16. Thank you TALEB BAHMED Ismehen #ai #computational #artificialintelligence #civilengineering #geotechnicalengineering #geomechnics #soilmechanics #soiltesting https://lnkd.in/dF_SXc9a

📃Scientific paper: Road Extraction with Satellite Images and Partial Road Maps Abstract: Road extraction is a process of automatically generating road maps mainly from satellite images. Existing models all target to generate roads from the scratch despite that a large quantity of road maps, though incomplete, are publicly available (e.g. those from OpenStreetMap) and can help with road extraction. In this paper, we propose to conduct road extraction based on satellite images and partial road maps, which is new. We then propose a two-branch Partial to Complete Network (P2CNet) for the task, which has two prominent components: Gated Self-Attention Module (GSAM) and Missing Part (MP) loss. GSAM leverages a channel-wise self-attention module and a gate module to capture long-range semantics, filter out useless information, and better fuse the features from two branches. MP loss is derived from the partial road maps, trying to give more attention to the road pixels that do not exist in partial road maps. Extensive experiments are conducted to demonstrate the effectiveness of our model, e.g. P2CNet achieves state-of-the-art performance with the IoU scores of 70.71% and 75.52%, respectively, on the SpaceNet and OSM datasets. ;Comment: This paper has been accepted by IEEE Transactions on Geoscience and Remote Sensing Discover the rest of the scientific article on es/iode ➡️https://etcse.fr/nLWOF