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