Automated colony counting on agar plates remains a bottleneck in microbiology workflows, with manual counting prone to fatigue-driven error and existing automated systems often requiring fixed imaging setups. We present a distance-invariant colony counting system built on classical computer vision, designed to operate reliably under variable camera positioning and backlit plate illumination. The pipeline detects the plate boundary via Hough circle transform, flattens illumination gradients, and applies watershed segmentation to separate touching colonies before contour-based morphometric filtering rejects non-colony artifacts. A two-layer anomaly detection system flags abnormal colonies: a statistical layer operating from the first plate, and a machine learning classifier designed to activate as labelled data accumulates. The system is organized around organism-by-plate-type profiles spanning 39 organisms and 11 media types, allowing profile-aware interpretation of colony morphology. Central to the design is a human-in-the-loop validation workflow, where users confirm, flag, or correct automated counts through a web-based interface, generating the ground-truth dataset required for the supervised learning layer. We discuss the architecture, design rationale for prioritizing measurement invariance and validation infrastructure, and the system's readiness for biological validation in ongoing work.
As a first-generation college student and Ronald E. McNair Scholar pursuing a Bachelor’s in Mechanical Engineering at Southern University and A&M College, I am driven by a passion for innovation and research. My experience has prepared me to tackle complex challenges. I am open... Read More →
In the 80’s, Gunnar Carlsson introduced a conjecture pertaining to the singular homology of certain finite CW complexes. In 2018, Iyengar and Walker demonstrated that a related conjecture was false using tools from homological algebra and a result on maximal rank of multiplication maps between exterior algebras over vector spaces in a paper by Conca, Herbig, and Iyengar. This presentation will showcase an investigation of maximal rank of a generalized form of multiplication map not discussed in the prior papers. In particular, such a generalized form of multiplication map between exterior algebras is induced by a certain partition of a finite dimensional vector space’s basis. Although more counterexamples can be produced in this way, this research indicates why vector spaces of dimension 2n induce certain multiplication maps that are well-behaved, and why more research is needed for the case of vector spaces of dimension dn.
Topological Data Analysis (TDA) uses techniques from algebraic topology to interpret high-dimensional and noisy data. Both multimodal and multiview Machine Learning (ML) aim to understand complex, high-dimensional data as well via the integration of different data types and different views of the same data type respectively. While their combination seems natural, the methods for implementation, the intent for their fusion, and the data sets used are varied across disciplines. This presentation looks at the ways in which TDA is being fused with multimodal and multiview ML from the approach of a systematized scoping review. In the development of this project, the Kitchenham guidelines and the PRISMA checklist for scoping reviews were followed as closely as possible. Of note is that the limited personnel behind the project means that the guidelines nor the checklist were able to be followed perfectly thus the project is instead a systematized scoping review. The papers reviewed were decided based in part by the inclusion of at least one TDA tool, use of multimodal or multiview data, and implementation of a multimodal or multiview ML framework. Preliminary results compare disciplines involved in research, purposes for development, and the ML pipelines used.
