What's Changed
- Create an inference pipeline for lung segmentation by @sriva-e in #2
- Initialize DVC and upload lung segmentation models by @ViacheslavDanilov in #4
- Boundary extraction by @sriva-e in #5
- Combine metadata files by @sriva-e in #6
- Extract subset from the MIMIC dataset by @ViacheslavDanilov in #7
- Stack frontal and lateral images by @ViacheslavDanilov in #8
- Supervisely Dataset Conversion by @ViacheslavDanilov in #9
- COCO dataset creation by @ViacheslavDanilov in #10
- Update lung segmentation by @ViacheslavDanilov in #11
- Visualize and validate COCO dataset by @ViacheslavDanilov in #13
- Implement pre-commits by @ViacheslavDanilov in #12
- Readme by @ViacheslavDanilov in #14
- Change the project structure by @ViacheslavDanilov in #15
- MLFlow integration by @ViacheslavDanilov in #16
- Integrate prototype model by @mak-en in #17
- Update MLFlow tracking by @ViacheslavDanilov in #18
- Fix box center coordinates by @ViacheslavDanilov in #19
- Update environment by @ViacheslavDanilov in #20
- Update SLY-to-INT converter by @ViacheslavDanilov in #21
- Continuous Integration by @ViacheslavDanilov in #22
- Hydra implementation by @ViacheslavDanilov in #24
- Implement DVC pipeline by @ViacheslavDanilov in #25
- Dry run different models by @ryndovaira in #23
- Fix floating point coordinates by @ryndovaira in #27
- Reset annotation ID when class ID does not exist by @ryndovaira in #28
- Implement inference pipeline by @ryndovaira in #29
- Upload labeled datasets to S3/DVC by @ViacheslavDanilov in #32
- Implement a pipeline for custom augmentation by @ryndovaira in #30
- Fix conf_threshold assignment by @ryndovaira in #31
- Visualization pm model by @mak-en in #34
- Add parameters to PM config by @mak-en in #36
- Update datasets and their metadata by @ViacheslavDanilov in #33
- Modify training pipeline by @ViacheslavDanilov in #35
- Visualize and evaluate detections by @ryndovaira in #37
- Bugfix/remove lateral images by @mak-en in #39
- Anchor generator by @ryndovaira in #38
- Set batch size for the detection pipeline by @ryndovaira in #41
- Updating img_scale in all pipelines by @ryndovaira in #42
- Implement additional backbones (ResNet-50 and VGG-16) by @mak-en in #43
- Implement Non Maximum Suppression by @ViacheslavDanilov in #40
- Implement Faster R-CNN by @ViacheslavDanilov in #45
- Implement model evaluation using threshold ranges by @ViacheslavDanilov in #46
- Implement ResNet-50 encoder by @mak-en in #47
- Edema pipeline by @ViacheslavDanilov in #48
- Exclude heart, artery and bronchus from COCO dataset by @ViacheslavDanilov in #51
- Implement box fuser by @mak-en in #52
- Upload Faster RCNN models to DVC by @ViacheslavDanilov in #53
- Update training dataset and pipeline by @ViacheslavDanilov in #54
- Fix SLY dataset by @ViacheslavDanilov in #56
- Implement IoU threshold by @mak-en in #57
- Fix the prob_maps bug in MapFuser by @mak-en in #58
- Split subsets in the interim dataset by @ViacheslavDanilov in #59
- Visualize pipeline results by @ryndovaira in #60
- Update DVC datasets by @ViacheslavDanilov in #62
- Implement classification evaluator by @mak-en in #61
- Fix the coordinate shift of the ground truth bboxes by @mak-en in #63
- Feature detection models by @ViacheslavDanilov in #64
- Detection evaluator implementation by @ViacheslavDanilov in #65
- Reimplement predict pipline by @mak-en in #66
- Train single-class models by @ViacheslavDanilov in #67
- Upgrade visualization by @ViacheslavDanilov in #68
- Update project readme by @ViacheslavDanilov in #69
New Contributors
- @sriva-e made their first contribution in #2
- @ViacheslavDanilov made their first contribution in #4
Full Changelog: https://github.com/quantori/edema-quantification/commits/1.0.0