Updated 2023 (markdown)

Vadim Pisarevsky 2023-09-06 17:46:56 +03:00
parent 2e4551f79f
commit 2cc0abe20b

45
2023.md

@ -32,6 +32,51 @@
<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><> <><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
</pre> </pre>
## 2023-09-06
* Vadim:
* GSoC 2023 is over. Out of 12 projects 10 are successful. PRs have been submitted. Now the students should send the final report and video.
* [OpenCV China] Yuantao almost finalized GEMM patch. ARMv7 part still needs to be finished.
* [OpenCV China] Wanli will work on extending OpenCV DNN perf tests that should help to avoid performance regressions in future releases.
* Shiqi Yu (OpenCV China):
* OpenCV China is going to launch 1-week course on OpenCV. It will contain some basic information about OpenCV; plus, it may be collaboration with various companies who produce NPUs, cameras etc. The second part of the course may introduce concrete products, how to use them with OpenCV. So, besides education it will let us to build better connection with industry.
* The course is going to offline, not online (there are already tons of online courses). The first run is expected to be in October; ~50 participants are expected.
* Vincent:
* working on fuzzer-based tests; found one issue in image warping.
* would be useful to refactor core headers and, in general, reduce opencv build time
* opencv zoo looks interesting, why don't we advertise it more actively
* Vladimir D:
* extending public CI.
* Gary: OpenCV summer of code student who works on RISC-V acceleration of OpenCV could continue work if it adds more value. OpenCV will pay for this extra time.
* Dmitry:
* need to discuss object tracking; probably have some code to contribute to OpenCV on multiple object tracking (this is unrelated to bytetrack).
* Gary:
* just got back from a long vacation.
* Satya is preparing proposal on attracting more funding.
* basic framework for NERFs has been created by GSoC and other students. Thinking of where to put it, how to package it. Talking to ʎLabs.
* Alexander S (opencv.ai):
* Done:
* Alexander Panov:
* Merged PRs fix refineDetectedMarkers[ #24139](https://github.com/opencv/opencv/pull/24139),
* Extended QR code benchmark to use new Aruco-based algorithm with command line [#5](https://github.com/opencv/opencv_benchmarks/pull/5)
* fixed issue “Incorrect dictionary size returned from cv::aruco::extendDictionary [#24222] (https://github.com/opencv/opencv/issues/24222)”, opened [PR 24228](https://github.com/opencv/opencv/pull/24228)
* checked and commented issue QRCodes decoding without Quirc [#24225](https://github.com/opencv/opencv/issues/24225)
* Working on PR for charuco benchmark
* Abduragim Shtanchaev:
* [Finished first iteration](https://github.com/opencv/opencv/pull/24037) of the einsum layer about 10 days ago.
* Working on corner cases of it, namely reduce operation. See full [TODO](https://github.com/opencv/opencv/pull/24037) list here
* Alexander Smorkalov:
* Reviewed, tested, helped to resolve issues with PRs ## 24128, 24227, 24211, 24139, 24214, 24209, 24216, 24221, 23607, 24215, 24199, 24196, 24204, 24180. Issues #24142 and 24185 are in progress.
* Identified regression in the test system (TS module). Reverted related patch in #24223.
* Works on infographics on OpenCV optimization approaches (HAL, Universal Intrinsics, OpenVX, OpenCL, etc).
* Alexander Lyulkov:
* Fixed issue with MJPEG encoder. [Pull request](https://github.com/opencv/opencv/pull/24209) is merged.
* Working on converting torch7 (not PyTorch) dnn samples/models to ONNX.
<pre>
<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
</pre>
## 2023-08-30 ## 2023-08-30
* Vadim: * Vadim:
* GSoC 2023 final evaluation started. * GSoC 2023 final evaluation started.