Dear all, Here is some news from team members:
International Common training of skin datasets then common reporting. Both teams will proceed the following way: -Scope (Shuang): Skin diseases of the PASSION project + Skin Cancer + Psoriasis -Separate Datasets (Shuang): Changsha: Chinese data then open-source/black data. Basel: Swiss data then open source/black data -Timeline (Xiaoyu): next week: training of data; following week: first results -Visual Algorithm and Framework for deep learning will be identical for both teams. -Transfer method: CycloGAN (China: yellow skin to black skin; Switzerland: white skin to black skin) Discussion with IFD / ILDS 1. update on the PASSION project -Scaleup of project potential includes NTDs so importance of pushing NTD agenda forward (there is now a subsection of NTDs at the WHO dealing on dermatological aspects of NTDs). 2. NTDs in Zanizibar -there is a NTD distribution system 3. Conference planning -The RDTC annual meeting should be a hybrid platform virtual real and we should brainstorm about making the networking relevant - value of having a world class conference in dermatology coming from RDTC. Involvement of ILDS, WCD ? Conferences and competitions: Fuzhou smart initiative (October) and RDTC Tanzania (January): add slides on technical presentation to podcast by Philippe / Herizo Madagascar: Onsite Dataset collection begun last week (lockdown restrictions lifted) -Images stored on local server onsite. Access to data on the server for training. Tanzania (Zanzibar): -Around a total of 45 cases collected currently. -Scabies is going up in incidence in the past 2 weeks. There is still difficulty to access through a pharmacy but there is a distribution system for ivermectin: big benefits to be able to add this to the project in Zanzibar. China (Changsha): 1. A total of 400 sets of data have been sorted out this week, according to the discussion mode, others are still being sorted out. 2. 136 new cases were collected this week, including 50 cases in the dermoscopy room, 61 cases in the operating room and 25 cases in the provincial children. 3.Quality control: 10 sets of quality control data are randomly checked this week, among which 10 sets are qualified and 0 sets are not qualified, and have been modified. Up to now, there are 10 sets of qualified data and 0 sets of unqualified data. Problems: 1) When sorting out multiple choice questions in the table, the Numbers should be separated by ", "; 2/ For those lacking text information in part of the data, the basic information may need to be followed up according to the pictures.
Join us on our Instagram page
Comments