crosstyan/RapidPoseTriangulation
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
153 Commits 3 Branches 0 Tags
651b806a3ac980e43da92ee5d4141e06922509e5
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Daniel 651b806a3a Small updates. 2025-02-11 11:26:58 +01:00
Daniel 2c994eca44 Allow keypoint detections outside the image. 2025-02-11 11:26:33 +01:00
Daniel 24d706d030 Some small updates. 2025-02-05 16:29:08 +01:00
Daniel 1fb730ff55 Create custom ros message instead of using json strings. 2025-02-05 12:11:22 +01:00
Daniel c1cc6054d1 Some small updates. 2025-02-05 11:58:01 +01:00
Daniel b286efeb86 Updating jetson installation parts. 2025-02-05 10:33:02 +01:00
Daniel 3eb7696e31 Some small updates. 2025-02-03 17:19:02 +01:00
Daniel b41380b033 Eval with all panoptic cameras. 2025-01-28 15:30:23 +01:00
Daniel 321aa0a714 Increased 2D pose merging threshold. 2025-01-28 14:54:01 +01:00
Daniel 1cc71cce57 Dropping more old python code. 2025-01-28 11:50:50 +01:00
Daniel a674811fef Update readme. 2025-01-27 18:32:55 +01:00
Daniel 56718838f2 Improved bad results debugging. 2025-01-27 16:49:47 +01:00
Daniel 4b0d3a545e Updated results. 2025-01-27 16:47:34 +01:00
Daniel d8beaecc56 Changed detector for wb poses. 2025-01-27 15:46:55 +01:00
Daniel 45d003c2e9 Moved onnx build to main dockerfile. 2025-01-27 15:19:19 +01:00
Daniel f9127e9a40 Merge close 2D poses. 2025-01-27 14:20:52 +01:00
Daniel d886d1db7c Fixed time calculation. 2025-01-27 14:18:08 +01:00
Daniel a1f85a5a53 Updated results. 2025-01-27 12:13:37 +01:00
Daniel 9ce51f90b3 Renamed ros packages. 2025-01-21 18:11:49 +01:00
Daniel 9778c75bf0 Drop old code. 2025-01-21 18:01:14 +01:00
Daniel 2ba9e096ac Updated image callback to directly calculate poses. 2025-01-21 17:15:33 +01:00
Daniel 29c072400f Mixed updates and improvements. 2025-01-21 16:23:19 +01:00
Daniel c5f190ab35 Eval skelda datasets with cpp implementation. 2025-01-21 15:10:43 +01:00
Daniel d77fee7103 Restructuring some code. 2025-01-20 18:00:37 +01:00
Daniel a866485c8e Small updates. 2025-01-20 16:38:19 +01:00
Daniel 86d8ccf797 Batching of poses in cpp implementation. 2025-01-20 12:20:25 +01:00
Daniel 8f2322694a Merge model outputs directly in graph. 2025-01-17 18:33:19 +01:00
Daniel 97ea039f7d Some code speedups. 2025-01-17 18:02:17 +01:00
Daniel 99368e5216 Reimplemented wrapper in cpp. 2025-01-17 16:42:05 +01:00
Daniel 8a249a2f16 Small improvements and fixes. 2025-01-17 16:24:37 +01:00
Daniel 1bd58deede Created ros wrapper for 2D pose estimation. 2025-01-15 16:08:24 +01:00
Daniel aec70d0418 Updated other results. 2025-01-15 13:47:01 +01:00
Daniel 0411837279 Implemented resize before debayering. 2025-01-15 12:20:10 +01:00
Daniel b804ce149d Testing bayer decoding on gpu. 2025-01-14 17:56:36 +01:00
Daniel 2f52521b9a Using bayer encoding for images. 2025-01-13 16:50:08 +01:00
Daniel ba234628eb Small docker fix. 2025-01-13 14:43:26 +01:00
Daniel 5187310d16 Update of skelda submodule. 2024-12-19 16:20:21 +01:00
Daniel 0dd0f84e19 Updated results. 2024-12-19 16:02:18 +01:00
Daniel b483269697 Renamed new wrapper script. 2024-12-19 14:23:41 +01:00
Daniel 7b073249d6 Deleted old 2D pose wrapper. 2024-12-19 14:22:48 +01:00
Daniel 5bc7bd9094 Splitting 2D and 3D predictions into single loops to prevent CPU priority issues. 2024-12-19 14:13:12 +01:00
Daniel db47a8af6a Updated dockerfile. 2024-12-19 14:01:07 +01:00
Daniel 2ad20549f2 Small updates. 2024-12-19 13:52:44 +01:00
Daniel 19bb1a6372 Use fp32 model for whole-body poses. 2024-12-19 11:12:13 +01:00
Daniel 1bb7dd966e Converted wb model as well. 2024-12-18 16:53:32 +01:00
Daniel f8984f9408 Some mixed updates. 2024-12-18 16:28:13 +01:00
Daniel 07426fac2f Optional batched pose processing. 2024-12-18 16:22:08 +01:00
Daniel 7b8d209601 Cache onnx-trt models, they are faster than using two trt-engines. Removed trt-runtime again. 2024-12-18 14:28:25 +01:00
Daniel b26ec998b3 Using tensorrt runtime directly. 2024-12-18 12:10:45 +01:00
Daniel 6e8f6a22ba Running mmdeploy benchmark. 2024-12-18 11:51:05 +01:00
Daniel 21e93661f5 Updated dataset configs and results. 2024-12-16 14:43:54 +01:00
Daniel 17ff41010a Some small updates. 2024-12-16 14:34:44 +01:00
Daniel 07f75c53ee Renamed a parameter. 2024-12-09 14:17:35 +01:00
Daniel 23108dd594 Further small improvements. 2024-12-06 18:15:08 +01:00
Daniel ee8b9bafb3 Updated easypose scripts. 2024-12-06 17:35:49 +01:00
Daniel 108937d96c Some mixed improvements. 2024-12-06 13:58:34 +01:00
Daniel 7a253cd615 Moved image cropping out of the graph again. 2024-12-06 12:32:51 +01:00
Daniel acf1d19b64 Improved box cutting with always fixed tensor shapes. 2024-12-04 17:54:57 +01:00
Daniel 6452d20ec8 Some small updates. 2024-12-04 11:46:06 +01:00
Daniel 97ff32b9ce Moved pose pre/post-processing into onnx graph. 2024-12-03 17:59:31 +01:00
Daniel 2e5c3399ed Workaround problem of tensorrt hanging sometimes. 2024-12-03 17:39:45 +01:00
Daniel 742d2386c7 Moved detector pre/post-processing into onnx graph. 2024-12-03 11:52:55 +01:00
Daniel 36781e616b Use rgb input for both models. 2024-12-02 17:28:34 +01:00
Daniel dc44a71b2c Some further speedups. 2024-12-02 17:00:04 +01:00
Daniel c8c48e4bf2 Using TensorRT backend. 2024-12-02 13:50:35 +01:00
Daniel bf837fbca9 Adding first results. 2024-12-02 13:46:17 +01:00
Daniel b44d4134d9 Move fp16 casting into onnx models. 2024-12-02 13:45:26 +01:00
Daniel 9cc63532d3 Extracted onnx tools into a single script. 2024-12-02 12:32:53 +01:00
Daniel d50364cda7 Some further optimizations. 2024-11-29 18:22:48 +01:00
Daniel ad1ef42b99 Export pose postprocessing steps as well. 2024-11-29 17:59:15 +01:00
Daniel 19015c6326 Exporting fp16 onnx models. 2024-11-29 17:37:44 +01:00
Daniel 7b9505ca02 Moved image normalization into onnx models. 2024-11-29 17:05:43 +01:00
Daniel 1b5e0c44e3 Speed up preprocessing. 2024-11-29 16:19:06 +01:00
Daniel 93d4611a91 Fixed running custom onnx models. 2024-11-29 15:18:57 +01:00
Daniel f6d13ea5a7 Testing onnx runtime with easypose. 2024-11-29 14:25:57 +01:00
Daniel ae21a3cf29 Updating base image. 2024-11-29 13:53:47 +01:00
Daniel 4b2811c1c8 Exporting onnx instructions. 2024-11-29 13:51:20 +01:00
Daniel 1c0eeb9147 Visualizing intermediate steps. 2024-11-15 15:05:25 +01:00
Daniel caa4a0cb51 Evaluated chi3d dataset. 2024-11-15 15:00:24 +01:00
Daniel 3b69bcfa61 Changed docker base image. 2024-10-29 16:41:43 +01:00
Daniel f7e7fb4e91 Small updates. 2024-10-29 15:00:03 +01:00
Daniel 9e03533704 Renamed project. 2024-10-14 16:28:34 +02:00
Daniel 256fd26f3f Updated skelda version. 2024-10-14 16:10:19 +02:00
Daniel 87a205c36c Update readme and h3wb results 2024-10-10 18:25:56 +02:00
Daniel e8986e0b2b Use more joints for final merge. 2024-10-10 18:04:29 +02:00
Daniel c3783835dd Update results. 2024-10-10 16:56:07 +02:00
Daniel b958bc8bcf Reset with index jumps. 2024-10-10 12:47:41 +02:00
Daniel b80ad139e5 Improved group merging. 2024-10-10 12:46:11 +02:00
Daniel 2f093e9f82 Evaluation of egohumans dataset. 2024-10-09 10:40:53 +02:00
Daniel fa2af31349 Improved group average calculation. 2024-10-08 18:11:46 +02:00
Daniel 99aa3bb301 Implemented fisheye camera triangulation. 2024-10-08 15:11:38 +02:00
Daniel 3318afc3f1 Delete old py code. 2024-09-26 12:10:05 +02:00
Daniel d2efb66767 Added flag for minimal group size. 2024-09-26 11:55:43 +02:00
Daniel f62d90f753 Updated readme example images. 2024-09-25 15:55:18 +02:00
Daniel 21f0a20d79 Keep with few limbs. 2024-09-25 15:42:06 +02:00
Daniel 42ac2731f1 Print duration stats. 2024-09-25 15:26:43 +02:00
Daniel cbf3526f26 Small fixes and updating other results. 2024-09-25 15:21:38 +02:00
Daniel ba910e11e6 Some fixes. Updated benchmarks. 2024-09-17 18:23:41 +02:00
Daniel 8eba0ffa65 Converted postprocessing steps to cpp as well. 2024-09-17 16:10:25 +02:00
Daniel 75840cf045 Better compiler optimizations. 2024-09-17 13:04:37 +02:00