A how-to talk on getting started in computer vision and data annotation with the the San Francisco Bay Association for Computing Machinery group
I recently had the pleasure of being invited to speak about getting started with computer vision, data annotation, and object detection with the San Francisco Bay ACM (Association for Computing Machinery) group.
I’d like to give a huge shoutout to all the organizers for running such a smooth virtual event! I encourage you to check out their upcoming talks!
For upcoming machine learning workshops check out sixgill.com/events
Watch the Talk & Find Resources Below:
Build your own object detection model from start to finish. Includes overview of computer vision applications, step-by-step instructions on data annotation, and model training with your own dataset.
Sense supports labeling with rectangles (bounding boxes) for object detection, polygons (masks) for instance segmentation, feature points for pose estimation, classes for classification, and a generic text option
Sense also supports most major export formats such as COCO, VOC Pascal, Create ML, YOLO, and Sense JSON.
Detectron2 Google Colab Notebook:
In this workshop we use the Detectron2 library from Facebook AI. It’s a great way to get started building models for object detection, instance segmentation or pose estimation! You can find links to my notebook used during the talk, and the official resources below:
- Google Colab notebook used during the workshop
- Official Detectron2 github
- Detectron2 model zoo
- Official Detectron2 documentation
- Official Detectron2 Colab Notebook
Still have questions? feel free to reach out to me, or join an upcoming live computer vision workshop!
If you build something with what you’ve learned we’d love to see it!
If you’re looking for managed labeling, custom model building, or enterprise solutions contact Sixgill to see how we can help.
We do multiple free live workshops almost every month! Checkout sixgill.com/events for more information.