A schistosomiasis dataset with bright- and darkfield images
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Schistosomiasis is a neglected tropical disease (NTD) that threatens 700 million and impacts 250 million people per year.The disease is caused by blood flukes of the genus Schistosoma, which enter the human body through contact with infected water.One species, S. haematobium, sheds eggs through the urinary tract, and can thus be diagnosed by examining urine samples for these eggs.Because concentrations of schistosomiasis infection are highly localized and are often in remote areas, rapid and robust field diagnosis is crucial to both individual diagnosis and the mapping that informs control efforts.Artificial intelligence (AI) algorithms, if properly designed, can speed up and improve both diagnosis and mapping through scalable, accurate analysis of images of urine samples.To develop such algorithms, we offer the dataset described here.It consists of paired bright-and darkfield images of urine samples collected in two distinct field studies in Cte d'Ivoire, Africa.There are images from 728 patients, of whom 151 were schisto-positive and contain S. haematobium eggs.Crucially, each patient has sufficient images to diagnose S. haematobium infection, so the dataset can be used to realistically test the diagnostic value of algorithms for clinical use.The division into two studies allows testing of algorithm generalizability.Due to exigencies of the data collection protocol, the images display a variety of qualities, from clear to blurry, which further allows testing of algorithm robustness to realistic noise.The dataset is thus well-suited to developing algorithms that can be of concrete value in schistosomiasis control efforts.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it