A methodology for morphological feature extraction and unsupervised cell classification
Bibliographic record
Abstract
Abstract Cell morphology is an important indicator of cell state, function, stage of development, and fate in both normal and pathological conditions. Cell shape is among key indicators used by pathologists to identify abnormalities or malignancies. With rapid advancements in the speed and amount of biological data acquisition, including images and movies of cells, computer-assisted identification and analysis of images becomes essential. Here, we report on techniques for recognition of cells in microscopic images and automated cell shape classification. We illustrate how our unsupervised machine-learning-based approach can be used to classify distinct cell shapes from a large number of microscopic images. Technical Abstract We develop a methodology to segment cells from microscopy images and compute quantitative descriptors that characterize their morphology. Using unsupervised techniques for dimensionality reduction and density-based clustering, we perform label-free cell shape classification. Cells are identified with minimal user input using mathematical morphology and region-growing segmentation methods. Physical quantities describing cell shape and size (including area, perimeter, Feret diameters, etc.) are computed along with other features including shape factors and Hu’s image moments. Correlated features are combined to obtain a low-dimensional (2-D or 3-D) embedding of data points corresponding to individual segmented cell shapes. Finally, a hierarchical density-based clustering algorithm (HDBSCAN) is used to classify cells. We compare cell classification results obtained from different combinations of features to identify a feature set that delivers optimum classification performance for our test data consisting of phase-contrast microscopy images of a pancreatic-cancer cell line, MIA PaCa-2.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".