Version 0.6: tensorflow support and colourmaps

This release adds support for Tensorflow models like Faster-RCNN and Mobilenet-SSD. It also gives you the option of applying a colourmap to single channel images. You can switch between CPU and OpenCL/OpenCL when running classification on images.

Version 0.5: batch inference and better progress bars

You can now run batch inference (CPU) over your project.

Discovered QProgressBar. Oops. Also improves some "previous" folder handling so you can load new data with fewer clicks.

Deeplabel version 0.4 release (inference!)

Deeplabel now supports loading Darknet-based object detectors to annotate images. Sample models trained on MS-COCO can also be downloaded from here for experimentation.

• Adds model inference via OpenCV DNN
• Supports Yolov3, Yolov3-spp, Yolov3-tiny (etc)
• Use your existing .cfg, weight and .names files

v0.3 Release

Adds test support for COCO annotation export, and support for loading videos. Also some visual sprucing up (icons) and of course, bugfixes.

Deeplabel version 0.2 release

This release offers a number of general bugfixes and includes support for image zooming and Pascal VOC export.

v0.1 Release

Initial Windows and Mac binaries for evaluation purposes.