Cancer Diagnostic Ultrasound Imaging with Deep Learning (CDUDL) Workshop

On 7 October, the Cancer Diagnostic Ultrasound Imaging with Deep Learning (CDUDL) Workshop, organised by Ten-D Buckingham Centre, allowed researchers and students to discuss the recent development in ultrasound imaging.

The number of ultrasound imaging research for cancer diagnostic is growing rapidly, and so are the deep learning technologies. The CDUDL workshop aims to be a forum for deep learning researchers and clinicians to present the application of deep learning based automatic ultrasound image analysis in cancer recognition, detection, segmentation and understanding. It also aims to have a two-way knowledge exchange between a radiologist and computer scientist.

Chaired by Dr Alaa AlZoubi and Mr Hongbo Du, introducing the Ten-D Buckingham Centre history and presenting the current achievements of the centre with the future plans for the team. The workshop focus was on the advance topics in deep learning for ultrasound image analysis and automatic cancer recognition.

Our guest speaker, Doctor Yi-Cheng Zhu from Renmin Hospital, presented keynotes about how doctors use the 2D ultrasound image to visually detect tumour and identify signs for different types of cancers such as thyroid, breast and ovarian.

The researchers of Ten-D Buckingham Centre had the opportunity to share their knowledge and experience on their own research while also learning more about the progress of the team. With Professor Sabah Jassim presenting the involvement of Persistent Homology in Deep Learning for Medical Image Analysis through the identification of distinctive characteristics. This led to the involvement of Topological Data Analysis by Jehan Ghafuri and how it can be used in medical imaging. With tumours being identified through object detection using advance deep learning methods, this allows the increase in accuracy which PhD student, Anu Bose, presented to the team.

Each member of the Ten-D Buckingham Research and Development Centre gave an insight into their progress and future plans to better their research. With opportunities for students to attend workshops such as this becoming more frequent at the School of Computing, this encourages their development and prepares them for external conferences.

Dr Alaa AlZoubi said, “CDUDL workshop provided the perfect opportunity for knowledge sharing between the medical practitioners, mathematicians and computer scientists. The way that oncologist extract visual information from the ultrasound images helped our team to identify the challenges and the development of our ongoing research”.

PhD research student, Jehan Ghafuri said, “CDUDL workshop was an excellent opportunity for us to learn about ultrasound images from an experienced specialist’s perspective. Dr. Zhu explained different tumour types, and challenges with ultrasound images.”