Updated 2020 (markdown)

2020.md

@@ -30,6 +30,52 @@
 <><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
 </pre>
 
+## 2020-08-26
+
+### *__Minutes__*
+
+1. Vadim:
+   - GSoC: the final evaluation for students has started. Need to ask the students to submit the final evaluation form.
+   - GSoC: Julia bindings project is 100% finished; the final video was received. Article for the blog is in progress.
+   - GSoC: RANSAC project is 100% finished. The final video was received. Article for the blog is in progress.
+   - Discussed DNN project evolution for OpenCV 5. There probably be no radical refactoring of the module in OpenCV 5 (e.g. towards NGraph-like architecture), however there are going to be some important improvements:
+       - reworked importers, especially ONNX importer.
+       - an external utility that will do careful network diagnostic and probably some transformations to simplify its use within OpenCV.
+       - probably an updated layer fusion part that will do multi-pass fusion.
+       - python samples could automatically download and store in some "cache" missing deep learning models.
+       - even better CUDA/cuDNN support, maybe via dynamically loaded "plugin", similar to opencv_ffmpeg, dynamically loaded FFMPEG wrapper plugin on Windows.
+   - Continued discussion about 3D module. Rostislav from Intel team will transform 3D module evolution proposal and will put it to OpenCV wiki
+2. Anna:
+   - travelling
+3. Gary:
+   - Asked Ajit to prepare the final request to expose some internals of Macbeth detector. Here are some results: https://github.com/opencv/opencv_contrib/pull/2644.
+   - April tag detector is in progress. Will it be finished? Gary suggested to re-use some of existing implementations.
+4. Alexander:
+   - initial support for SYCL (OpenCL One API). It will be put to OpenCV 4.5.0
+   - "next" branch for 5.0 will be created.
+   - function to draw OpenCV logo has been added
+   - audio input/video support is in progress. Windows & linux are covered.
+   - GSoC:
+     - SIFT PR is under review. video report has been prepared.
+     - WASM optimization is done. Performance data is being collected.
+     - Depth fusion is on track. Students is polishing the final PR.
+     - Object tracking: PR has been merged.
+   - OSPP:
+     - QR encoder/decoder is generally working pretty good. The student will spend the rest of time improving QR encoder.
+5. Shiqi:
+   - GSoC: text detection & recognition, Wenqing submitted updated PRs with faster networks. Video is still to be improved.
+   - Found 2 potential developers for OpenCV, who can ready contribute to OpenCV part time.
+   - Found some more potential members of OpenCV development partnership program. Negotiations are still going on. This year we might get up to 3-4 partners in total.
+   - OpenCV 5.0 proposal: one company that produces web cameras with depth sensors; they are interested to participate in 3D module improvements. Ethernet cameras support is also very important: actively used in autonomous driving and in medical areas.
+   - Continue work to find more partners for OpenCV, in particular, hardware companies.
+6. Stefano:
+   - some more discussion about use of some standard compiler stack for OpenCV DNN
+   - object tracking pull request is merged into 3.4. Blog post can be prepared.
+
+<pre>
+<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
+</pre>
+
 ## 2020-08-19
 
 ### *__Minutes__*