MMAI Project

I planned to work on the IRMA dataset.

The dataset consist of more than 10,000 medical images, with tags in predefined format for each image. The dataset is widely used in the training and evaluation for medical image relative systems.

Part of the dataset is used in the imageCLEF medical retrival and annotation task, which is the most well known image retrival competition.

I planned to build a system that can perform classification task on the dataset. The system can be extended to carry out other tasks such as automatic diagnosis if other data can be obtained.

It can also be extended to carry out the imageCLEF medical retrival task, which allows the system to utilize text knowledge. Since imageCLEF2011 dataset will not be available untill Februrary, 2011, I will focus on the IRMA dataset first.

Group Member:
B95401011 蘇昱銓