|
|
--- |
|
|
license: mit |
|
|
--- |
|
|
|
|
|
### Dataset Description |
|
|
We created a training dataset comprising 11,178 oocysts collected from 231 images from midguts dissected at day 7–9 post infection. |
|
|
Using this dataset we trained a machine learning model to count and size oocysts in mercurochrome-stained mosquito midguts. |
|
|
See the oocyst counting/sizing tool [here](http://oocystmeter.org/) |
|
|
Preprint coming soon |
|
|
|
|
|
### Dataset summary |
|
|
This dataset includes 231 panoramic images of mosquito midguts imaged at 100x or 200x magnification. |
|
|
The midguts were stained with mercurochrome to reveal <i>Plasmodium falciparum</i> oocysts. |
|
|
Oocysts were manually labeled with the VGG Image Annotator (VIA) using ellipses tracing their outline as drawn by trained parasitologists. |
|
|
We also labeled the outline of each midgut using the “polygon region shape” function. |
|
|
|
|
|
The annotations were saved in json format and split into the same train and test split used to train the OocystMeter model. |
|
|
|
|
|
|
|
|
### Species |
|
|
<i>Anopheles gambiae</i> mosquito midgut infected with <i>Plasmodium falciparum</i> oocysts |
|
|
|
|
|
|
