v2.2.0

CVHub520/X-AnyLabelingv2.2.0Dec 26, 2023by CVHub520

AI Summary

X-AnyLabeling v2.2.0 is a significant update to this annotation tool, introducing major UI/UX improvements including a new status bar, background color rendering, and expanded editing capabilities. The release adds robust import/export support for multiple annotation formats and enhances the Segment Anything (SAM) annotation workflow.

Key Highlights

Extended rectangle box editing to four-point editing mode for more flexible annotations
Added comprehensive import/export support for YOLO/VOC/COCO/DOTA/MASK/MOT label formats
Introduced hide/show selected objects feature to address layer overlap issues
Enhanced SAM annotation mode with one-click clear point prompts
Added real-time status bar showing mouse coordinates and object dimensions

Breaking Changes

Removed confirmation dialog for object deletion - deletions are now immediate
GPU version compiled with CUDA 11.6 - may cause compatibility issues with different CUDA versions

New Features

Label background color rendering
SAM annotation mode with one-click clear point prompts
Four-point editing mode for rectangles
Auto-switch to editing mode when cursor hovers over objects
One-click export for segmented mask images
Import/export for YOLO/VOC/COCO/DOTA/MASK/MOT formats
Data statistics for current task
Hide/show selected objects
Real-time preview of filename and annotation progress
Support for 'Difficult' label annotation
Bottom status bar with mouse coordinates and object dimensions
Direct modification of object text descriptions in label editing box

Full Release Notes

## Released versions

| Version | MD5                                |
|---------|------------------------------------|
| X-AnyLabeling-CPU.exe | 9036a66c1badc298f3e565eed261a67e |
| X-AnyLabeling-GPU.exe | 94441d8f32fdbbdd1293bae88e723b0f |
| X-Anylabeling-Linux-CPU | 8bfbe8112bfd5c97a8daf02635c7fd8d |
| X-Anylabeling-Linux-GPU | 10ee27a6ed960a08edb3eb37c2d8b06a |

> Note: The GPU version mentioned above is compiled based on CUDA 11.6. Different versions may result in issues with xxx functionality. It is recommended to consider running the application using the source code, especially if compatibility issues arise.

## Supported models

| Task                     | Model                  |
|--------------------------|------------------------|
| **Image Classification** | ResNet50, InternImage, YOLOv5-cls, YOLOv8-cls, PULC Person/Vehicle Attribute |
| **Object Detection**     | YOLOv5, YOLOv6, YOLOv7, YOLOv8, YOLOX, YOLO-NAS, DAMO_YOLO, GOLD_YOLO, RT-DETR |
| **Instance Segmentation**| YOLOv5-seg, YOLOv8-seg |
| **Keypoint Detection**   | DWPose, YOLOv6-face, YOLOv8-pose |
| **Oriented Object Detection** | YOLOv5_obb        |
| **Multi-Object Tracking** | OC-Sort, ByteTrack   |
| **Segment Anything**     | SAM, SAM-HQ, EdgeSAM, MobileSAM, EfficientViT-SAM, Med-SAM2D |
| **Optical Character Recognition**                  | PPOCRv4                |
| **Land Detection**       | CLRNet                 |
| **Image Captioning**     | RAM                    |
| **Visual Language Model**| Grounding-DINO         |

## Important Updates

- ✅ Added label background color rendering.
- ✅ In SAM annotation mode, added the ability to one-click clear point prompts.
- ✅ Extended rectangle box editing mode to four-point editing mode.
- ✅ Added support for automatically switching to editing mode when the cursor moves over an object.
- ✅ Removed confirmation dialog for object deletion.
- ✅ Implemented one-click export function for segmented mask images.
- ✅ Added one-click import and export for YOLO/VOC/COCO/DOTA/MASK/MOT labels.
- ✅ Introduced data statistics for the current task.
- ✅ Newly added functionality to hide/show selected objects, addressing layer overlap issues.
- ✅ Real-time preview of the current filename and annotation progress.
- ✅ Added support for "Difficult" label to annotate challenging samples.
- ✅ Added a bottom status bar to display real-time mouse coordinates, as well as the width and height of the selected object.
- ✅ Enabled direct modification of object text descriptions in the label editing box.