Titulo Estágio
Trustworthy model for image-based cancer diagnosis
Local do Estágio
LARN-CISUC
Enquadramento
## Internship
Recent advancements in artificial intelligence have improved our lives by allowing us to make better decisions and to be more efficient and have better results.
An area which has benefitted greatly by these advancements is Healthcare. Nowadays, doctors are able to use advanced models using Deep Neural Network that provide very accurate results. However, in Healthcare, more than in other areas, trustworthiness and accountability are necessary. A doctor needs to understand why a program made a certain diagnosis so he/she can make an informed decision. And, if the prediction is wrong, it is necessary to understand why the program made that decision.
A particular case where trustworthiness is important is in the detection of cancer. In this situation, not only is it fundamental to be able to predict if a person is going to have cancer, but due to the long-term bad effects the treatment has on its patients, it is necessary to be able to dismiss the possibility of it occurring. This program will provide useful explanations for a doctor while at the same time will be at the same level of accuracy as the state-of-the-art models.
Objetivo
## Objectives
The internship aims to propose, develop and test a trustworthy, accurate and scalable model for cancer diagnosis using histology images (microscopic images of biological tissues). To achieve this purpose there exist several datasets, e.g. for cancer detection using histology images:
- C_NMC_2019 (https://wiki.cancerimagingarchive.net/display/Public/C_NMC_2019+Dataset%3A+ALL+Challenge+dataset+of+ISBI+2019) with 15,135 images from 118 patients (cancer or leukemia)
- Textures in colorectal cancer histology (https://zenodo.org/record/53169#.YMACFzoo-V5) with 625 images of 8 classes of tissue (both Cancer and Non-Cancer)
Plano de Trabalhos - Semestre 1
### 1st Semester
- Literature review (State-of-the-art-models,Image preprocessing and feature engineering, Current techniques for explainability, Evaluation metrics)
- Identification and study of available frameworks
- Analyse and define datasets and scenarios
- Define the architecture of the system
- Start implementing the proposed approach
- Write intermediate report
Plano de Trabalhos - Semestre 2
### 2nd Semester
- Implement the proposed solution and fine tune models
- Test and evaluate performance
- Write final report
Condições
## Conditions
A Grant will be available during the second semester internship depending on the 1st semester candidate evaluation. A 3-month scholarship of 835,98 euros per month is foreseen for this work, renewable for another 3 months. Logistics @Laboratory of Neural Networks (LARN) DEI-FCTUC
Observações
N/A
Orientador
Bernadete Ribeiro; Catarina Silva
bribeiro@dei.uc.pt 📩