Esta é unha revisión antiga do documento!
Downloadable Research Results
This is a repository of datasets, software and algorihtms.
The construction of this repository is an on-going process. Most of the content included here is owned by CiTIUS researchers. Please read carefully every page for detailed authoring and usage restrictions.
This work is licensed under a Creative Commons Attribution 4.0 International License.
You can also reach these pages by using one of these URL aliases:
- http://wiki.citius.usc.es/software/<name>
- http://wiki.citius.usc.es/datasets/<name>
- http://wiki.citius.usc.es/algorithms/<name>
Algorithms
- GPU-Retina - GPU-Based Segmentation of Retinal Blood Vessels.
- Split-and-Merge - GPU Algorithms Based on the Split-And-Merge Method.
- GPU-ELM-remoteSensing - GPU-Based Extreme Learning Machine for the Classification of Hyperspectral Remote Sensing Images.
- ELM-EMP - Classification of Hyperspectral Images using ELM and EMP.
Classification Datasets
- Fish Ovary - 4 sets wih a total of 3.868 instances, 25 (rep. 41) inputs, 2 (resp. 3) classes to detect the cell nucleous (resp. to classify maturity cell states: hydrated, cortical alveoli and vitellogenic/atretic). The inputs are texture features extracted from histological images of fish ovary.
- Scab - 126 instances, 256 inputs, 2 classes (presence/absence of common scab in potatoes). The inputs are bands of infrared hyperspectral imaging data of potatoes.
- Pollen grains - 291 instances, 23 inputs, 3 classes (species of plants Parietaria Judaica, urtica urens and urtica membranacea). The inputs are shape and texture features extracted from microscopical images of pollen grains in the air.
- Honeybee Pollen - 2.600 instances, 70 inputs, 5 classes (species of plants cytisus, castanea, quercus, rubus and paphanus). The inputs are texture features extracted from images of honeybee pollen loads captured by a stereomicroscope conected to a digital camera.
- CITIUS-USC Video DB - Video and eye fixation's database with 72 videos observed by 22 subjects. It contains 40.558 fixations aquired with the SMI RED eyeTracker (The videos include four categories, natural and synthetic scenes, with fixed or movement camera). The fixations have been acquired for both eyes while the subjects stand on a 1.280×1.024 monitor located at 60 cm of distance.